>In defining what it means for a mathematical structure to be of conscious experience, our
proposal does not answer the question of what this mathematical structure is.
Nothing is answered in this article. We simply don't know.
This article is up here because nobody is sure whether or not chatGPT is in any way representative of an aspect of our conscious experience.
Is chatGPT a mathematical model of consciousness?
We know it doesn't represent the whole experience because... it has huge flaws... but we aren't sure whether it represents an aspect of what we are.
Disagree or agree, we have something in our hands and it could be it... but at this point we don't know.
> We know it doesn't represent the whole experience because... it has huge flaws
To be clear, we can't yet conclude that these huge flaws are due to its structure rather than the extra sensory data that we have and it doesn't. Doesn't really change your overall point, just wanted to clarify what we can and cannot conclude from the evidence, because people are far too quick to jump to conclusions after seeing the flaws.
Completely agreed. There are flaws. But there are times where the output appears that it cannot anything else other then complete understanding of what was given to it as input.
The vast majority of information cannot be compressed and the proof of this is trivial by a simple counting argument.
In any case, it is still just a lookup table. Compression, unless lossy, doesn't change the meaning of the information in this lookup table. It just makes access more indirect.
Hutter might think that intelligence is a form of compression, but I think that's just wishful thinking. It might be my ego speaking to say I think I'm the one that's right. But it's not my ego speaking to say that there's no hard proof to say unequivocally that intelligence is equivalent to compression.
> The vast majority of information cannot be compressed and the proof of this is trivial by a simple counting argument
It doesn't have to be lossless compression. Intelligence is also about being discerning about the information that's relevant, although that's basically the same point.
> But it's not my ego speaking to say that there's no hard proof to say unequivocally that intelligence is equivalent to compression.
Actually I think it's kind of trivial to see that intelligence is a necessarily form of compression. Abstractions, concepts, theories, these are all lossy descriptions that discards irrelevant details and capture only relevant details to make problem solving via deduction, induction or abduction tractable. And yet, using those concepts, theories, abstractions you can reproduce/predict observations with very high fidelity. That's compression.
Do you have a counterexample of something that we consider intelligence that doesn't fit this model? I suppose it depends what specifically you mean by "equivalent".
>Nothing is answered in this article. We simply don't know.
No, we haven't clearly defined consciousness. For a given exact definition of consciousness, we can determine whether something fits that definition or not. If we don't clearly define it, then we're just talking about a feeling, a qualia, and it's not possible for people to reach a consensus because they're not even discussing the exact same thing when they say "consciousness".
The consistency arises from the fact that even though we don't know the definition, we are in agreement in what is classified as conscious.
We agree that humans are conscious, we agree that rocks are not conscious. This classification is clear and points to a underlying definition we cannot fully articulate. It is more than a qualia or whatever you call it. We know it has exact parameters simply because we are able to perform classification tasks.
The consensus breaks for things like chatGPT. chatGPT straddles some uncertain border that we can't articulate. But in general for the term consciousness we reach consensus by how we classify things as conscious or unconscious.
> The consistency arises from the fact that even though we don't know the definition, we are in agreement in what is classified as conscious.
Clearly there's not agreement in what is classified as conscious, given this thread/article.
>We agree that humans are conscious, we agree that rocks are not conscious. This classification is clear and points to a underlying definition we cannot fully articulate. It is more than a qualia or whatever you call it.
This is a fallacy. People agree that yellow is not violet and green is not violet, but they don't agree that indigo is not violet. Just because there are some classifications on which people agree about the label this doesn't imply they agree upon all classifications. A more direct example is whether plants are conscious or not.
>Clearly there's not agreement in what is classified as conscious, given this thread/article.
There is general agreement. We agree for most things like rocks and humans..there are a few things where we aren't sure, like chatGPT.
But in general humans are in agreement for most things.
>Just because there are some classifications on which people agree about the label this doesn't imply they agree upon all classifications.
Well just because ambiguities for certain things exist doesn't preclude the existence of things that are utterly clear to be conscious or unconscious.
We know humans are conscious and we know humans have all the required parameters to make them classified as conscious. We just can't deduce the exact thresholds and features that cause a phase shift when changed. But we know these parameters exist.
> Is chatGPT a mathematical model of consciousness?
Unless we reduce models of consciousness to purely stochastic processes, this is trivially false. And no one seriously reduces consciousness to probabilistic pattern recognition.
> We know it doesn't represent the whole experience...
It doesn't represent any experience. ChatGPT doesn't "experience" things as conscious things do. It's not even comparable to the experience of a simple invertebrate, for example.
That's the (hard) problem of qualia. You can't say for sure that ChatGPT doesn't experience anything. And I can't say for sure that you do, either. It's trivial to dismiss it on the basis of "well obviously it's like that"—far more difficult is to say something about why it's not like that.
I think consciousness has to be the product of an algorithmic process. I think this process can be summed up as "prediction". I don't think there's some secret quantum/soul/magic aspect to it. Which is why I think it's plausible that transformer models may capture some aspect of what we refer to as consciousness, though I don't quite buy it at a gut level.
You say that "it's not even comparable to the experience of a simple invertebrate," but that's just you writing down what your gut is telling you. What is the experience of a simple invertebrate like, exactly?
>You can't say for sure that ChatGPT doesn't experience anything. And I can't say for sure that you do, either.
The question rattling around in my brain tonight is - if some, many, people don't develop a theory of mind as young children, how many of the people around me are like this? Is it just the philosophers?
It seems of much more moment than the essence of ChatGPT.
Given how many people in this thread don’t seem to grasp the idea of qualia, and think the hard problem of consciousness to be easy, I’m not entirely unconvinced I’m surrounded by a bunch of imperfect p-zombies.
>Unless we reduce models of consciousness to purely stochastic processes, this is trivially false. And no one seriously reduces consciousness to probabilistic pattern recognition.
Most things in this universe have mathematical models that can describe it. We have models for virtually everything. Thus it is reasonable to say such a model also exists for consciousness. To say what you said definitively without even knowing the meaning of the word consciousness is premature. You don't have enough knowledge to say anything about this.
>It doesn't represent any experience. ChatGPT doesn't "experience" things as conscious things do. It's not even comparable to the experience of a simple invertebrate, for example.
Again you don't know what you're talking about. Nobody can define what consciousness is and nobody fully understands the emergent effects of chatGPT or the human brain. To say what you say with conviction is therefore empty. You don't actually know because nobody knows.
I think that's a bit uncharitable. You're acting like people haven't been grappling with these problems for thousands of years. Searle, for example, makes a pretty good argument why symbol-processing systems will never produce consciousness (you'd have to prove that the Chinese room is conscious, etc.).
It's clear, for example, that a toaster doesn't have any conscious experiences, but a dog does. To say that "nobody fully understands" the human brain is throwing out the baby with the bathwater. ChatGPT is not conscious any more than my (very complex) Swiss wristwatch is.
Philosophers grapple with the problem the same way Greeks used mythology to grapple with the problem of reality. Thunder is not thrown by some god named Zeus. and the impossibility of a mathematical model describing consciousness cannot be explained away by a Chinese room argument that has tons and tons of qualitative holes in it.
If an atom can be described by a mathematical equation that means so can groups of atoms. If groups of atoms can be described math and if a neuron is simply a complex group of atoms then a neuron can be described by an equation. If neurons can be described by an equation then so can groups of neurons. If the brain is a group of neurons then the brain can be described by a mathematical equation. If consciousness is a physical phenomena produced by a mathematical brain then consciousness is a mathematical model.
The only way you can believe that consciousness cannot ever by described by math is if you see some break in the previous chain of logic. Most likely it's in the linking of consciousness entirely with the physical brain. Likely you believe in a soul or something if you don't agree with the logic. Which just means you're using mythology and concepts outside of science to derive your conclusions. Again, lighting is not thrown down from the sky by a god named Zeus.
It's clear a toaster is not conscious. But what chatGPT is, is not clear. Placing it in a category where we aren't sure is reasonable and also it's a category agreed upon by many people. It would be a bit uncharitable to demand that your categorization of chatGPT is more correct then other peoples. Even you must be be aware of the ambiguity associated with chatGPT despite your stubborn stance to the contrary.
I would posit that it is clear. It's simply a whole bunch of toasters put together (that is, a long list of if-then statements). If you're trying to argue that consciousness is merely a bunch of if-then statements, then yes, ChatGPT might be conscious. But then it's going to be much harder to argue that a toaster isn't conscious, which I thought we both agreed with.
> Likely you believe in a soul or something if you don't agree with the logic.
This seems a bit dismissive, and although I myself am a substance dualist (which is generally seen as a metaphysical position), most materialists aren't mechanistic in their interpretation of consciousness.
>I would posit that it is clear. It's simply a whole bunch of toasters put together (that is, a long list of if-then statements). If you're trying to argue that consciousness is merely a bunch of if-then statements, then yes, ChatGPT might be conscious. But then it's going to be much harder to argue that a toaster isn't conscious, which I thought we both agreed with.
The combination of multitudes of if-else statements forms a macro structure. It is this emergent macro structure that we call consciousness. As stated before the exact interconnect of this macro structure can't be fully articulated yet.
>This seems a bit dismissive, and although I myself am a substance dualist (which is generally seen as a metaphysical position), most materialists aren't mechanistic in their interpretation of consciousness
My arguments are based on likelihood derived from logic and evidence meaning While we don't fully know if a soul fits in this framework, we certainly do not yet have the capability to derive it's existence/likelihood from evidence or first principles. Thus there's really nothing to talk about as it's an argument outside of known logic. What is the point of arguing about topics we don't know about? There is no point. I'm being dismissive because there is literally nothing left to talk about for either of us. Logic and evidence is also the only shared principles that seem consistent across individuals so best keep discussions in that realm fully.
> The combination of multitudes of if-else statements forms a macro structure. It is this emergent macro structure that we call consciousness. As stated before the exact interconnect of this macro structure can't be fully articulated yet.
It can be, and it's called the theory of computation. It's been extensively studied for the past century and a half, and, all-in-all, it's not even particularly strong. For example, most real numbers aren't even computable, Godel showed that consistent theories can't even prove their own consistency, and so on. In my view, our mind does not seem to compute things (though some, like Dennett and Putnam, would disagree).
The claim that our mind computes things is a pretty strong one (and if correct, I think we'd all probably be much better at math). Penrose argues that whatever the mind does, it's (at least in part) a fundamentally unsound operation; this to me sounds way more correct than the idea that our mind operates algorithmically.
>It can be, and it's called the theory of computation.
It's the most likely outcome given what we know of the physical world.
> Godel showed that consistent theories can't even prove their own consistency, and so on.
Godel showed a flaw within logic itself. But it's all we got. If you're operating outside of logic then we don't have a consistent logical framework to use for generating conclusions. Without this, like I said, the conversation is pretty much over. There is nothing left to say.
Can you explain how you know that neither a toaster, your wristwatch nor ChatGPT are conscious?
If being conscious means being conscious like a human this would be obvious but maybe the toaster is just conscious like a toaster.
Is a bug conscious?
Internally we know we're conscious because we just do. We experience qualia, even though we don't know what "we", "experience", and "qualia" really mean.
Even so. For us, consciousness is just there.
Externally we attribute consciousness to certain behaviours, including (but not limited to) self-reference, goal-seeking, physical and social responsiveness, use of language, memory, and so on.
These two things are not the same. We assume other humans are conscious in the same way we are to the extent that they perform consciousness. We can't be sure, but for human interactions it's a workable proxy.
A toaster, a watch, a rock, and a web server don't perform consciousness in any way at all. They may have some kind of metaphyiscal awareness, or not, but if they do we can't see it, and without a comprehensive theory of consciousness it's parsimonious to assume they're not aware.
ChatGPT performs some elements of consciousness, but not others. It happens to perform the elements AI researchers consider important - specifically use of language. But it doesn't have a memory for previous conversations, it doesn't show any independent goal seeking, and so on.
It uses "I..." and a vast amount of human training to mimic the appearance of consciousness, and plenty of humans seem to want to believe that's enough.
But the essence of the hard problem is the difference between "performs the actions we associate with consciousness" and "is conscious in the same way we know we are."
You could argue that practically there's no difference, and from one point of view that may be true. But that doesn't work if you want to understand what's happening rather than just assuming in a "well obviously..." kind of a way.
If only because historically in science "well obviously..." has been consistently wrong.
They may have some kind of metaphyiscal awareness, or not
Another option is there is some kind of metaphyiscal awareness everywhere, and it's only coincidental that objects of class "human" eventually convince themselves into believing that it's equivalent or highly related to their own thinking process. Unless we exclude this option somehow, AI can't be definitely seen as simply rolex, because it does that too.
Currently AI is discontinuous and passive, but is that even required?
without a comprehensive theory of consciousness it's parsimonious to assume they're not aware
This hypothesis creates more entities (one island of awareness per complex enough object), so why not the other way round?
It's clear, for example, that a toaster doesn't have any conscious experiences
So anthropic. We can agree that it is on the opposite end of a spectrum though. A spectrum that we define out of a single point (“us”, and mostly “I”). We can also say that a toaster is a part of your consciousness instead, like e.g. your arm but less noisy. But then you are a part of a dog’s consciousness as well.
Consciousness is probably best described as just a field in a complexity space, not a discrete set of localized phenomenons. It has a near-zero value in a physical projection of a toaster’s concept, but is that projection even useful?
Consciousness apparently arose from evolution, which is a funny process of optimization under stochastic conditions. (There's also a good amount of pattern recognition involved.)
Why is it that optimizing for survival can lead to consciousness, but not predictive coding? Nothing here is trivial.
While some people believe that consciousness cannot be reduced to probabilistic pattern recognition, I strongly disagree. In my view, the AIXI model is an excellent model of intelligence, as it can mathematically define a notion of Pareto optimal behavior across all possible environments. This model relies on compression at its core, specifically the minimum description length principle.
Moreover, large language models (LLMs), like GPT-3, are designed to learn how to predict, and prediction is the dual of compression. As explained in this StackExchange post (https://cs.stackexchange.com/questions/130655/is-prediction-...), prediction can be used to compress data, and compression can be used to make predictions. Given that LLMs are trained on vast amounts of data and can predict patterns within that data, I believe that they contain all the necessary components to be considered an intelligent system.
Therefore, I firmly believe that consciousness can indeed be reduced to probabilistic pattern recognition, and that LLMs provide a compelling example of this phenomenon in action.
Your viewpoint is one I consistently have to push against.
Computation is not necessarily experience. I can't emphasize enough how fundamental understanding this is. It seems some just can't grasp this concept.
Would you say a program that echoes to the screen the responses of a hidden human is intelligent? Viewed as a black box, it would certainly seem so. But does it have any internal experience and understanding of anything? Of course this is a trivial experiment, but the point of it is to show, in the most absurdly clear terms, that producing plausible outputs !== being intelligent, much less having a conscious experience.
This example also shows that the perceived sophistication of the input-output transformation is immaterial, unless you want to say consciousness requires a certain internal architecture for computation, in which case I would ask: what architecture is that? No matter the sophistication of an algorithm, it's computation lies on a different axis to that of consciousness. You might say human compute because they are conscious, not conscious because they compute.
Another way to understand this: Let's say you touch a very hot object. You skin, in conjunction with your brain, senses and computes the temperature of the object. But, there is no essential reason why that should lead to an experience of pain, rather than a sensation similar to color, which generally does not produce pain.
> This example also shows that the perceived sophistication of the input-output transformation is immaterial
No, it doesn't show that, it merely asserts it. Consciousness could be a logical property of some perceptual algorithm, like ordered outputs is a property of a sorting algorithm.
> But, there is no essential reason why that should lead to an experience of pain, rather than a sensation similar to color, which generally does not produce pain.
Sure there is, because it's approaching the point of causing damage. Pain happens when sensory precepts approach acceptable tolerance. Pain is like clipping in an audio stream.
Except you didn't. The black box you described does have feelings because the human inside has feelings. The fact that every component of a sentient system doesn't have feelings isn't particularly interesting.
> There are plenty of damaging processes in the body that produce no pain.
Because they didn't need to for fitness purposes.
> What I’m interested in is how sensations are produced;
> Because they didn't need to for fitness purposes.
Huh, damaging process are not relevant for fitness purposes? Wow
Your link is full of speculation, and indeed for consciousness, speculation may be all that can be done, since no one can define it, only experience it.
> You don't know that.
It seems self-evident. If a program computes the temperature of the sun, does it feel the heat?
> Huh, damaging process are not relevant for fitness purposes? Wow
Cancer or heart disease in your 50s does not impact your reproductive fitness, for instance. Way to attach a strawman.
> It seems self-evident. If a program computes the temperature of the sun, does it feel the heat?
The sun isn't computing a temperature, it's computing a nuclear fusion reaction. Also, "feeling" isn't a quality of individual particles to the best of our knowledge, but it is a quality of some aggregate of particles we call a human.
So if your question is whether a faithful simulation of a human would feel the heat from a faithful simulation of the sun, then yes, that's very plausible.
> Cancer or heart disease in your 50s does not impact your reproductive fitness, for instance. Way to attach a strawman
But there are plenty of cancers that can and do occur in ones reproductive years, that start without pain.
> The sun isn't computing a temperature, it's computing a nuclear fusion reaction.
I wasn’t comparing the program to the sun. I was comparing it to the human skin.
> Also, "feeling" isn't a quality of individual particles to the best of our knowledge, but it is a quality of some aggregate of particles we call a human.
Hmmm, now you are referring to particles and aggregates. Wonder what the source of that claim is. At least you agree computation (a process) is not the important thing, but particles (matter).
> So if your question is whether a faithful simulation of a human would feel the heat from a faithful simulation of the sun, then yes, that's very plausible.
Wow, now you're making a circular argument. Whats a “faithful” simulation of a human? How would we know it is “faithful”? What form would that even take?
Let’s say ChatGPT becomes “sufficiently” advanced, and is hooked up to a sensor that records the temperature of the sun. Would it feel the heat? The whole idea is preposterous.
> But there are plenty of cancers that can and do occur in ones reproductive years, that start without pain.
So? The incidence of these cancers is insufficient to impact reproductive fitness in our species. At no point did I claim that all damaging processes must produce pain, I only claimed that all pain originates from evolutionary pressures indicating high chance of damage that does impact fitness. That also does not mean that presence of pain is always directly indicative of damage, because biology is messy and precision has a cost to fitness.
> At least you agree computation (a process) is not the important thing, but particles (matter).
Nope, not what I'm saying at all. The particles are a computation in a simulation, so the information content of the computation is what matters.
> Wow, now you're making a circular argument. Whats a “faithful” simulation of a human? How would we know it is “faithful”? What form would that even take?
A faithful simulation is clearly a simulation that contains all of the information content of a human and replicates the mathematical relationships that govern the operation of a human's mind. Since a human is a bounded volume, it contains finite information per the Bekenstein Bound, and is thus amenable to simulation, unless you are asserting some kind of non-physical dualism.
It seems clear that you don't yet grasp the argument being made so maybe you should reserve judgment, like calling the argument "circular", until you do properly grasp it.
> Let’s say ChatGPT becomes “sufficiently” advanced, and is hooked up to a sensor that records the temperature of the sun. Would it feel the heat?
It wouldn't feel heat in the same way that humans feel heat because its sensor and information processing is not mathematically equivalent to human information processing and sensory systems. That does not mean it wouldn't feel something like its own version of heat. A whole lot of vagueness hides behind your "sufficiently advanced".
If ChatGPT v10.0 did reproduce the mathematical relationships underlying human information processing and sensory systems, then yes, it would feel heat in precisely the same way as humans do.
"It seems some just can't grasp this concept.". Your confidence reveals your lack of knowledge. No one has a scooby what consciousness is or how to measure it. It could be that every atom has a modicum of conscious experience. It could be that consciousness is an emergent phenomena (although I think this claim is unlikely as it seems to me to be making a category error). It could be that it's actually Midi-chlorians. I repeat nobody knows. There are as many theories as there are people who have thought deeply about the problem. You coming along all confident with your debunking mindset and your scientism are just showing everyone who has any understanding that you do not know what you are talking about.
To me, it is those claiming that computation equals consciousness who are displaying hubris.
Think about it, if we accept the theory that computation equals consciousness, then then is little to investigate further. Just calculate something, and voila, consciousness exists.
It is this closing down of inquiry that I push against. My assertion is that consciousness is deeper, it cannot be reduced to computation.
How would you even prove otherwise? If it arises from computation, what is conscious? Is it the algorithm? Is it the electric circuits? It is the physical machine?
It is interesting to note my consciousness is bound up what a notion of what my body is. I can identify the limits of my body and make a distinction between it and my environment. What determines the conscious boundary of a entity performing calculations?
On the other hand, there are many strong arguments against consciousness arising from mere computation, some of which I have enumerated.
> And no one seriously reduces consciousness to probabilistic pattern recognition.
All eliminative materialists disagree. To them, consciousness is mechanistic in some way, so why not stochastic pattern matching which can be made Turing complete quite easily. For instance: https://www.pnas.org/doi/10.1073/pnas.2116933119
> It doesn't represent any experience. ChatGPT doesn't "experience" things as conscious things do.
You don't know that because we don't have a mechanistic understanding of "experience".
I'm surprised that in all this hype Gödel Escher Bach hasn't seen a proportional comeback. Hofstadter outlines with some formality a very plausible kind of structure of consciousness, namely the "strange loop": a system of symbols or rule-patterns that contains a distinct but isomorphic (or approaching) description of itself.
The book doesn't really go that much into the hard problem of consciousness, as much of the ideas are orthogonal to that discussion. If Nagel asks, "What is like to be a bat?", Hofstadter asks "If the mind of a bat were a program, what characteristics would it have?"
IMO the hard problem is not posed to be solved. I think of it like this. The world in which we are p-zombies is categorically indistinguishable from the one in which we are not, by an objective perspective. It is not a concept that we should use for an objective definition.
The issue is, what else would we call the rock that does all the things a CPU does? If we can socket it into the motherboard and play games on it, what makes it so different?
In tandem with that article there are also all those findings about GPT training smaller gradient descent models inside itself which is very interesting. That particularly seems rather... strange.
I'm sure Hofstader would dislike people saying "he was right about this!" "he defined this!", though, since that's not what he was setting out to do at all. GEB really is just an exploration, with some of the results being couched in mathematical language as an artifact of his background, not as an attempt to formalize anything.
Personally I think Hofstadter was a bit too enamored with recursion. I don't think it's some marker of consciousness. I'm not saying chatGPT isnt an aspect of consciousness (I don't know).
I suggest you look into Roger Scruton's interpretation
of Wittgenstein's private language argument. I think it might go something like this:
You want an explanation for the sensation of the color red? Alright, I'll try, as long as you help me out: what exactly is the sensation of the color red? I need to have a good grasp on what we're talking about here.
If you're trying to suggest qualia don't exist or don't have anything to do with consciousness, then sure you'll find some people who agree. But it's not a mainstream position.
And of course the entire point is that qualia are ineffable. That's part of the definition. It's the feature, not a bug. But if you want to deny that you have conscious sensation then go ahead. ;)
At some point I figured that "everything has qualia" and "qualia doesn't exist", when taken to their logical conclusions, are actually identical beliefs. Though I suspect I'll need quite a lot more crack cocaine to figure out how.
I don't understand the mystery around "qualia". You sense something and it propagates through the system priming other neurons. It's just learned association.
What I'm saying is different from self awareness. Self awareness could arguably just be having a model of yourself. I'm talking about a sensory feedback loop.
Is there a book or resource that describes the state of the art in our knowledge and understanding about the nature of consciousness (Not cognition) and provides a survey of the popular and promising theories?
Readable by someone who isn't in the field(s). Maybe even a good popsci book on the topic that is the neuroscience's equivalent of "Universe in a nutshell"
For the "Universe in a nutshell" version overview of neuroscience, I'll plug my own book Journey of the Mind: How Thinking Emerged from Chaos (more on that below)
Unfortunately both miss what I think is easily the best and most computationally robust and fleshed out consciousness theory, but that's because it's so tragically unknown.
I co-authored a book that covers this (while staying accessible)
https://www.goodreads.com/book/show/58085266-journey-of-the-...
So much of the stuff in the mathematical-this and physics-that of conscious experience is just crap because there is very little understanding of biological consciousness and what it is and where we can find it, and where we cannot. These start with assumptions of what consciousness is and then go on to mathematize these non-validated assumptions. (Integrated Information Theory) IIT is the most popular among these. https://scottaaronson.blog/?p=1799 -- this article captures well why IIT, and many like it, are bad models of consciousness.
Thanks for the reply! I have to say, I am quite giddy. When I posted this question, I wasn't expecting to get a reply from an author and researcher on this very topic.
I am a software engineer by profession, but this is a topic that has been on the back of my mind for many years. Even when I read popsci books about the universe and the big bang and so on, i felt spiritually unsatisfied because they often do not address this elephant in the room - consciousness. A theory of everything that explains the laws of universe will be imo incomplete if it doesn't also explain, at the fundamental level, the laws of the universe that exists behind our eyes. It really deserves more attention and awareness among geeks and the general public.
You might be waiting a long time. Things like Tarski's Undefinability Theorem (moreso than Gödel) make it rather unlikely that this will turn out to be possible.
There has been a turn in consciousness research in the last in the decade towards some kind of panpsychism for precisely your sentiment: a completed physics should also be a completed phenomenology. David Chalmers has written extensively and seriously entertained panpsychism in recent years.
Galileo’s Error, published in 2019, by Phillip Goff traces the history of the mind-body problem and illustrates how we divorced our consciousness from the physical world during the development of early modern science and philosophy.
Last century, Ernst Mach and Bertrand Russell argued in favour for neutral monism, a stance which holds that the mental and physical are two sides of the same coin. It seems like these ideas are coming back after a long stretch of dogmatic physicalism.
Panpsychism gets a bad rep for sounding like it has associations with vacuous or flimsy New Age beliefs, but it means a totally different thing in the context of philosophy of mind and consciousness studies. It is the claim that consciousness might be a building block of the universe.
The main problem with panpsychism is not that it is incredulous, but that it currently lacks a compelling explanation for our individual subjectivity. If consciousness and mental properties are ubiquitous in the universe, why do I have my consciousness and you have yours? The main motivation to believe in panpsychism seems to be its incredible elegance and self-consistency. That alone does not make it true, but it might be a very big hint on where to look for answers.
I am sorry to disappoint but I think consciousness is the one thing that we will never be able to solve physically. It is perhaps the only thing that is true about that universe, that I somehow exists. Everything else is simply derived from that and might as well be an illusion (perhaps a very consistent and steady illusion).
> It really deserves more attention and awareness among geeks and the general public.
There's 0 interest in consciousness because it's so hard to pay attention and realize that it exists. We go about our days for weeks, months, years and even decades without even noticing that consciousness is a thing.
> We go about our days for weeks, months, years and even decades without even noticing that consciousness is a thing.
Do we? I feel like I'm constantly reflecting on it and trying to make sense of it, and that I'm certainly not alone. I'd like to think that it's part of being human to turn the lens back on itself and wonder "what am I?" But I suppose some people could care less for navel gazing.
This is a common view, but it is based on humans thinking they are special. “My own thoughts are so amazing”, “my own subjective reality is so real to me”, thus surely this is unexplainable.
When really, humans give themselves (their own thoughts) way too much credit and most are barely aware of their own thoughts, barely self-aware at all.
It is easy to say something is ‘un-explainable’ and thus ‘must be in some otherworldly realm that is beyond comprehension’, if one is not able to observer the problem. Most humans aren’t self-aware enough to know their own thoughts or examine where those thoughts come from.
Just a few decades ago simple voice recognition was deemed ‘a totally human ability that the computer will never accomplish’, now it is ordinary.
It comes down to humans internal subjective reality is both amazing and ‘unknowable to the outside’, ’un-observable’, thus un-testable. But then we make this giant leap that nothing we will ever build will also have some ‘internal subjective reality’.
It keeps coming up that GPT is a glorified auto-complete, but then never explain how a human isn’t also just an ‘auto-complete’ machine, just going through the day twitching at stimuli. Why do we think a human doing copy/paste, and regurgitating something they heard, is special, but if a computer does then suddenly it is completely invalid to think they could ever be conscious.
What you are referring to is the soft problem of consciousness, the contents of consciousness. You are probably right that the contents of human consciousness, or qualia, can be approximated by our understanding and technology to a very high degree given enough time.
However, OP was referring to the hard problem of consciousness. The hard problem of consciousness is unsolvable IMHO by science, as it is entirely subjective; in fact, it is the fact of existence of subjective experience. The problem is this: even if you find some objective mechanism that generates consciousness, how would you make the explanatory leap from the objective mechanism to the subjective? Even if you say: "Oh, we flipped this switch and now this person is unconscious." - this is a description from the outside. You cannot claim by this that the person's consciousness has disappeared. Maybe not connected to the contents of the human mind anymore; but the gulf between objective and subjective can never be crossed by objective science.
If I had a dollar for every time someone claimed that the hard problem of consciousness isn't actually hard, I would be very rich.
I think people have confused the matter by splitting into soft/hard.
There is only 1 consciousness. Soft is describing how it can be done physically, how to build it, or explain it, and hard is what it is like to experience it from the inside.
By creating the hard problem, it really just says "we can't know anything about any internal state". All "internal subject states" are self reported, and a machine can just as well self report enough to mimic a human. So at that point, we can't prove that the machine doesn't have a 'subjective' internal state.
By this line of reasoning, we can't prove that any humans are conscious either. Any lines of logic that try to prove the hard problem, would either both prove that the machine and human are both conscious, or prove that neither are conscious.
Sorry, you are part of the 90% confusing the hard and soft problem without understanding what is hard about the hard problem. In fact, merely realizing the hard problem of consciousness is one of the first stages that is reached in modern vipassana meditation.
The hard problem is not stating "we can't know anything about any internal state", it is about the mere subjective existence of the internal state and not about the contents of internal experience or any objective fact about that internal state. I don't know why this is seems often so hard, but in consciousness research and meditators' language (or in my native language at least), most of this seems pretty well-definded.
It has no relation to proving that humans or machines are conscious. It is about no objective fact at all, and thus cannot be investigated using objective methods. That is why it is unsolvable by objective science.
I was probably unclear. Meant, "we can't know 'others' internal state". We can examine our own, but all we can do is try to express it in language. Someone else hearing our description can only get a vague idea. And vice versa. Each person has their own individual internal subjective perspective, which can't be communicated with 100% clarity or accuracy. Since we can't examine or measure others internal state, we deem it 'hard' and 'impossible'. Because of course it is hard to express something that can't be communicated.
I would say that this examination of the problem of our internal view of the world, versus the reality of the world, subject versus object, goes back thousands of years.
I've also meditated a long time, and I do understand the hard problem. I just don't think eastern philosophies have the answer either. Eastern philosophies definitely had a jump start on the West with examining the mind, and really it was Buddhism that informed western philosophies from Schopenhauer onward. They helped defining some vocabulary and hierarchies, like outlining Skanda's or parts.
BUT, the typical eastern answer, that it is just 'this" and other 'wooo-wooo' answers are just as far from any explanation to the 'hard' problem as anybody in the west has. Meditation can get you awareness of the hard problem, some insight, but it still isn't an explanation. It just gets you another perspective that is still "Internal". I can meditate forever, but any 'realization' is still internal to me, it isn't providing any answers or explanation.
This comes up a lot right now with AI getting so human like (recent advancements), and I would say that at some point we'll create something that will be indistinguishable from a human, and at that time if the AI speaks convincingly of their realization of 'this is it', we wont have any ability to disprove that they really have had the same 'wooo-wooo' feelings. There is no way to prove anything.
So what I am speculating here, is that at some time in near future, we will be able to correlate soft and hard problems, so we have a map from one to another. And we'll have to realize that our own internal 'experience of reality' is just our neural net in our brain reacting and processing to stimuli, and we'll be able to examine the same things in silicon. And we'll have to accept that we aren't some super mystical beings or that our 'consciousness' is not all that special a development in the universe.
Or more to the point. In humans, our thoughts arise on their own. We don't think about what we are going to think about. The thoughts just appear. We are going through the day processing stimuli and reacting, thoughts popping into our head and leading us along. Our brain is just on a game loop, processing events. Except in biological humans, instead of math, it's calcium ions and potentials. Just an analog computer.
So lets say in current models', or future, in very large neural nets, that there will not be some internal subjective awareness, just like we have. We've mapped the entire neural network/nervous system, of a nematode worm and translated it into software. The 'software worm' appears to behaves the same as the real worm, so what is that software experiencing?
> It is perhaps the only thing that is true about that universe, that I somehow exists. Everything else is simply derived from that and might as well be an illusion (perhaps a very consistent and steady illusion).
Actually, there is every indication that you are an illusion.
Thanks! What you point out, the spiritual dissatisfaction, is one of big challenges for consciousness explanation. How can we ground this magical thing in the seemingly mechanical. How can that be satisfying? It's a tough one. But I didnt think that it would become a pressing one too so quickly with all the new AIs coming up and forcing us to think seriously about this question. What I come away with us we are the universe waking up to reflect on itself through the consciousness, self-awareness, and purpose imbued within us.
Also one of my first names :) I am kinda shocked I didnt actually make the connection until you pointed it out. I mean, I've always thought of Chaitanya as meaning enlightenment (one of its other meanings) and it's not like we periodically dwell on the meaning of our names.
Depending on what exactly you mean by “the nature of consciousness”, it’s fundamentally impossible to answer.
The fact that we think we can observe and comprehend a mechanism doing the observation does not preclude things we cannot see. You might be able to pan a cctv camera and get some insight about the camera itself if you find a reflection, but you can’t pan it over to the linux driver. It is not a given that all of the observer exists within the observable universe.
This school of thought strikes me as fundamentally opposed to the whole idea of the scientific method. We try to find out about these black boxes through repeated observation, theorizing, and testing.
It seems short sighted to suggest we can’t find out things about consciousness in a similar way.
The scientific method is about paying strict and close attention to what we can observe. The black boxes we can make sense of must have enough datapoints within the observable universe to be understood.
We can almost definitely learn a lot more about consciousness related phenomenon than we already know through science, and I don’t know the answer to OP’s question about the current cutting edge of that, but I think it’s always going to be about related phenomenon rather than consciousness itself. Unlike other phenomenon, which are only revealed to us because aspects of their existence appear in our observable universe, we don’t really “observe” the core experience of consciousness. The goalposts can be moved indefinitely whenever consciousness related phenomenon are made sense of. The fundamental experience is not observable, which means we can’t really see that black box at all. To me the mystery of that is much more satisfying than thinking we can ever get a definitive answer.
It is easy (and cognitively useful) to forget that, when the indirection is very reliable.
As far as consciousness, we won’t just have to speculate based on our own consciousness. As artificial minds become highly self-aware (a functional ingredient of consciousness), we will be able to construct very controlled experiments.
There will be endless numbers of experiments we will be able to do once we can edit self-aware entities and both externally query and internally monitor them.
True, but those experiments are inevitably restricted to the aspects of consciousness available for indirect observation and comparison.
There is no way to distinguish between an unconscious machine mimicking the functional aspects of consciousness we are able to indirectly observe and compare to our own experience and something having the same experience.
We could easily (and I think almost definitely are given the current pure symbolic manipulation track) be doing the same thing as a jungle cargo cult constructing an air strip out of sticks. Studying that won’t tell us how planes work.
Obviously advanced AI is more sophisticated than that and may have more legitimate parallels to consciousness, but I think it’s fundamentally impossible to know what’s a core part of the conscious experience and what just looks superficially similar to the output of conscious beings.
The current generation of AI are clearly not at all conscious and are “just” (don’t want to diminish the work) very good parrots generating image and text output from a given input in a way a whole bunch of conscious humans already have. But we’re already getting cargo cults and people who are anthropomorphizing and considering them comparable to humans.
There’s a danger in studying artificial models of an underlying phenomenon to learn about the phenomenon. That method of study is inherently very limited and extremely difficult to keep calibrated to the underlying phenomenon.
> True, but those experiments are inevitably restricted to the aspects of consciousness available for indirect observation and comparison.
As noted before, all observations are indirect. That has not been the unique challenge for studying consciousness.
The unique challenge has been an inability to create constructive or destructive experiments.
We didn't know how to construct consciousness candidates or precursors, of varying kinds to study. We are getting close to that (i.e. self-awareness as a functional quality is certainly a precursor.)
And we have not been able to create controlled demolition of an consciousness candidates, due to the morals of human experiments. (But our artificial minds will be alterable and destructible at will, to test hypothesis.)
You’re reiterating a point I acknowledged. There are subtleties you’re ignoring.
Our conscious experience is in fact a unique phenomenon because it is all perceivable phenomenon.
You can create as many self reflective machines as you like and experiment on them and compare them to output from humans, and maybe you’ll learn things, but you can’t ever get to looking at consciousness at all. There is no indirect or direct data you can use to compare all of your experience to something that sits inside a fraction of it.
What you’re talking about are behavioral experiments comparing the reported thoughts and actions of separate acknowledged conscious actors, ie other people, to the internal states and outputs of machines.
You’re not looking at consciousness when you do that. Looking at consciousness requires looking at all experience. Which is fundamentally impossible. When you try to look at all of experience as a single object to study nothing is differentiated. There is zero information.
The scientific method is part of an empirical philosophy and requires the ability to make objective measurements.
The problem with consciousness is that we have no ability to measure it objectively because it is an entirely subjective phenomenon.
You have no way of knowing if I am conscious or just a sufficiently elaborate machine doing a really great job of acting like I am (a p-zombie).
Personally, I lean towards panpsychism. I assume I am not special and that everything in the universe, or to put it another way the entire universe itself, has an experience of being.
Perhaps. But we don't know. Maybe a plausible theory exists waiting to be discovered. New findings in neuroscience and computer science give us tantalizing clues and hints about what may lie behind the curtain. This is a question that has haunted humanity for millenia and we are fortunate to be alive in the best time in history to ask it. Its not time to throw in the towel yet!
Of course not! Regardless of whether the core phenomenon is fundamentally impossible to explain (which I think it has to be for philosophical reasons), it’s an incredibly productive question to ask/am also curious about cutting edge theories and related findings.
Perhaps a model for explaining the model that explains the model and so on?, that is a recursive procedure to obtain a fixed point for the function f(model) = f(explain(model))? So that f = model0 where model0 is the fixed point of the function f. In order to use explain you should read first Kant's account of Reason (1). Hince knowledge should be expanded within the dialectics framework so that no concept of God or World emerge from it.
Psychedelics, schizophrenia and other mental illnesses, meditation and other techniques are reasonably valid examples of leaving I'd say. We don't really study them them to the degree that we could, but perhaps some day we will.
Psychedelics, meditation and mental illness show you or change the contents of the house - but you can never leave the house of the existence of consciousness. The closest thing we have to that would be death, and even then you only know if consciousness disappears if you die.
In fact, you would not even experience the disappearance of consciousness - because the consciousness that would experience this has disappeared. It is impossible to experience the absence of consciousness.
She has another book called Conversations on Consciousness that's basically transcribed interviews with prominent figures that's also worth checking out.
I can't say how good this book is, but the author has a PhD in parapsychology. This makes me question how prone they are to magical thinking, and also whether I'd want to read what they have to say about this topic.
Yes this comment is an ad hominem, and no I have not read the book... All I have is a skeptical hunch that this might not represent "the state of the art" on the topic.
After an initial, intense, personal experience understandably awakened her interest she always approached the subject from a scientific point of view:
From her wiki page:
It was just over thirty years ago that I had the dramatic out-of-body experience that convinced me of the reality of psychic phenomena and launched me on a crusade to show those closed-minded scientists that consciousness could reach beyond the body and that death was not the end. Just a few years of careful experiments changed all that. I found no psychic phenomena—only wishful thinking, self-deception, experimental error and, occasionally, fraud. I became a sceptic.
It is also a natural consequence of the culture you happen to have been raised in, which is an extremely interesting aspect of consciousness: how easily it can be programmed, and how influential it can be (it changes how reality "is"). In your case you have self-awareness of the phenomenon, at least in this case, but most people don't operate on that level.
It's understandable why you'd think this but it seems either didn't read much or were selective about her bio. Briefly:
* After spending time in research on parapsychology and the paranormal,[5] her attitude towards the field moved from belief to scepticism.[6][7] In 1987, Blackmore wrote that she had an out-of-body experience shortly after she began running the Oxford University Society for Psychical Research (OUSPR):[8][9]
"Within a few weeks I had not only learned a lot about the occult and the paranormal, but I had an experience that was to have a lasting effect on me—an out-of-body experience (OBE). It happened while I was wide awake, sitting talking to friends. It lasted about three hours and included everything from a typical "astral projection," complete with a silver cord and duplicate body, to free-floating flying, and finally to a mystical experience. It was clear to me that the doctrine of astral projection, with its astral bodies floating about on astral planes, was intellectually unsatisfactory. But to dismiss the experience as "just imagination" would be impossible without being dishonest about how it had felt at the time. It had felt quite real. Everything looked clear and vivid, and I was able to think and speak quite clearly."
In a New Scientist article in 2000, she again wrote of this:
"It was just over thirty years ago that I had the dramatic out-of-body experience that convinced me of the reality of psychic phenomena and launched me on a crusade to show those closed-minded scientists that consciousness could reach beyond the body and that death was not the end. Just a few years of careful experiments changed all that. I found no psychic phenomena—only wishful thinking, self-deception, experimental error and, occasionally, fraud. I became a sceptic.[10][11]"
* She is a Fellow of the Committee for Skeptical Inquiry (formerly CSICOP)[12] and in 1991, was awarded the CSICOP Distinguished Skeptic Award.[4]
* Susan Blackmore has made contributions to the field of memetics.[19] The term meme was coined by Richard Dawkins in his 1976 book The Selfish Gene. In his foreword to Blackmore's book The Meme Machine (1999), Dawkins said, "Any theory deserves to be given its best shot, and that is what Susan Blackmore has given the theory of the meme."[20] Other treatments of memes, that cite Blackmore, can be found in the works of Robert Aunger: The Electric Meme,[21] and Jonathan Whitty: A Memetic Paradigm of Project Management.[22]
This is all from her wikipedia, which is about a one minute read.
As I recall, her consciousness textbook is fairly opinionated but it still gives a fair tour of the mainstream theories. She happens to be an illusionist like Dan Dennett. There's no "woo" in her coverage of the topic.
You're asking the wrong question by excluding cognition and using a very ill-defined word ("consciousness"). "Consciousness" is not a thing. It's a collection of things, many of which are what we usually refer to as "cognition".
People who like to talk about consciousness have a problem keeping the goalposts in place (and trying to exclude cognition from consciousness, presumably because "cognition" doesn't sound mysterious enough, is a great example of this sort of moving of goalposts).
The truth is consciousness is an emergent phenomenon, and it can only be understood as such. Of course we have a lot more to learn, but we understand the fundamentals of neuroscience and biochemistry, and we know how our conscious experiences emerge from those. A neuroanatomy textbook (eg.: Blumenfeld) is probably the best place to start. There is unfortunately no single "thing" called "consciousness" that we will ever be able to point to.
That view seems pessimistic without any strong argument, other than critiquing people for not yet understanding something we all agree we don’t understand yet.
> Consciousness" is not a thing. It's a collection of things, […]
A collection of things is a thing.
> The truth is consciousness is an emergent phenomenon, […]
I don’t think that is controversial at all.
We have unraveled the workings of many emergent phenomena. Emergent doesn’t mean incomprehensible.
It's only a thing if you can define it, but there is no agreed upon definition of what collection of things comprise consciousness. We have all the pieces, and we know we have all the pieces (note: this doesn't mean we don't have a lot more to learn, both can be true). There is little controversy over how it all roughly works together. The only real controversy is semantics: what do we mean when we say "consciousness".
In programmerland, one might say something like "the abstraction you're using is counter-productive", for whatever reason, when looking at some junior hire's code. It's very much the same here: trying to understand consciousness is asking the wrong question. It's like asking "Why is there a God?": it's an unproductive question because it's a question that can't be answered. The abstraction in the question simply doesn't map 1:1 to abstractions in the world in which we live. If we want to understand our world, we'll maximize our odds of success if we focus on studying the abstractions of this world.
I wouldn't describe myself as a pessimist at all, and I'm not sure why you thought my comment was pessimistic... But this opinion of mine (because that's what it is, it's an opinion, not an argument) is a professional opinion. This isn't my field of expertise, but I do have expertise in very adjacent fields.
A phenomena can be defined by a pattern, even if it’s not well defined in a technical sense.
And a pattern is a thing.
Consciousness is an everyday experience that appears to be reliably shared by billions of entities who communicate about it to each other. And have done so for at least millennia.
There are types without values. That the type is itself a value does not mean that it has values itself. False is a type with no values. Consciousness is not a value.
Replace type with concept and value with thing, and we get a rebuttal using the same language. You can also switch the terms, depending on how 'a thing' is interpreted. If you are more familiar with set theory, replace type with proposition, value with element. The critical difference is between reference and referent.
Further if large numbers of people holding something is true for considerable duration is sufficient for something to 'be a thing', we'd also have an argument for 'the consciousness of the human heart' being a thing, which I don't think anyone would argue in good faith these days. There are probably better examples, but note that the duration of time and frequency of experience are parameters with no obvious values indicating at which point the underlying 'thing' should be accepted as undeniably 'a thing'.
The whole point of requiring it to be technically defined is that all this confusion just goes away and it becomes possible to ask more meaningful and less ambiguous questions about it.
That said, I will concede that the argument I have made does not say that consciousness is definitely not 'a thing', just that it is defensible to say that it is not a thing, given that it is held that either it does not represent a proposition, or that the proposition it represents is not true. (note I'm using the term proposition with some handwaving).
The most fundamental aspects of reality, and mathematics, are still not founded completely on formal definitions, without resorting to circular reasoning, and attempts to prove things assuming the things that are being proved.
Existence isn't well defined. Yet Descartes' famous informal claim "I think, therefore I am" is still a compelling statement that each of us can use to verify we do exist. Despite many unanswered questions.
Likewise, "I am aware of myself, my own thoughts, my interactions with a world around me" is a compelling statement supporting consciousness. The fact that we can all verify our own consciousness the same way, and discuss those experiences with each other, makes it even more compelling.
Despite lots of unanswered questions.
--
Your counter examples are also weak. Most subjective beliefs about hearts having consciousness are not nearly as credible as informal arguments for existence and consciousness.
Conscious hearts are not virtually universally experienced.
Informal reasoning about why a heart, a blood pumping organ, would correlate with consciousness are not generally accepted informal arguments.
People whose hearts are replaced with artificial or donated organs don't reliably report a loss of consciousness.
I concede that the subjective perception of consciousness is a thing, is a thing that I perceive, and that philosophy related to it is socially valuable work. I concede that there is no obviously correct choice of axioms for mathematics or logic. I concede that nothing is fully defined formally in that the formal definition will be of the conception of the concept and not of the thing which the concept might represent. I hold that the adequacy of a concept in representing something is an orthogonal concern to its formalization.
My claim is that while tools to determine if consciousness exists do not presently exist, it is possible that they are discovered, and applied, and that the question of the existence of consciousness is open. Note that for me existence is a fundamentally formal notion. I cannot defend this except on grounds of its anecdotal usefulness because it is akin to an axiom or ground truth.
The notion of consciousness I consider is with respect to its 'actual' existence[1]. Suppose I imagine a circle, and then a square, but believe I imagined at once a round square. That I can consider the notion of 'a round square' or may have had a direct perception of it does not require me to accept its existence. Likewise, I do not need to accept the existence of a unifying singular notion of what might be perceived as consciousnesses. This also does not necessitate I reject it. If consciousness cannot find an agreed upon definition, assertions regarding it lack semantic meaning and cannot be 'reasonably' (waves hands) assigned truth values. If an exact definition is determined with no logical contradictions, I defer to the physical sciences for determining its plausibility.
The notion of conscious hearts was more credible in the past. The idea was that before the role of the heart was understood, a large population could have been convinced that 'the heart is the seat of consciousness', and thus a large population believing they were experiencing something a particular way may have been followed by the falsification of the existence of such a thing. At the time, it was very plausible that it was considered impossible to ascertain whether or not there were thinking hearts. I concede that the notion is not presently credible; for my purposes it is sufficient that it was ever credible.
If you regard existence proofs as being extra-logical, it may not be the case we can find agreement since we will be effectively talking in different languages. Personally, I am very skeptical of the ability of the human mind to find truth informally, which is why I lean on formal notions heavily.
--
1: My favorite essay on this distinction is Quine's (admittedly polemic) "On What There Is". The distinction I consider is slightly different from his in that I am also expressing distrust of direct self-perceptions of the mind's functions.
Informal reasoning doesn't mean subjective reasoning.
Just because someone believes something, doesn't mean there is a compelling argument it exists.
But if something can be studied by many independent people, with strong agreement not based on passed-on beliefs, then there is likely to be something there.
I.e. I am unaware of a universal experience of souls, or informal arguments for souls that are compelling.
I am aware of informal arguments for existence ("I am, therefore I exist") and consciousness ("I experience myself, my own thinking, even this thought itself, so I am conscious of myself"). In both cases, the vast (vast, vast) majority of humans conclude they do exist and are conscious. And the informal reasoning is very strong, even informally tautological.
However, I find bringing the concept of "soul" into the equation to be helpful, because many people intuitively believe in such things. If you ask them if souls are important, they'll say yes. If you ask them if something special happens to souls after death, they might say yes. But ask them to define what they mean when they say "soul", and you'll get a ton of hand-waving. People can believe very strongly in things they can't even describe or point to: this always puzzles me, because it shows that people don't even know what it is they believe in in the first place.
Whether the concept of "consciousness" is as ridiculous as the concept of a "soul" or not, I have found that they both suffer from the same definition problem.
There are many, many concepts which throughout History entire societies have shared a belief in, but which didn't map to real things that could be defined precisely. A "soul" is one of those, but other examples abound in the History of science and in the History of medicine.
Who said anything about "magic"? That's straw-manning the literature on qualia. I'd like to hear your explanations for our sensations, because I'm not aware of any that actually explains them.
Consciousness is just your brain working. Everyone that tries to mystify or expand on it further is making shit up. The glasgow coma scale has existed for a long time in medicine. You literally lose parts of consciousness before the whole as various physiologic markers deteriorate. This isn’t even pessimistic. It’s just not an interesting explanation and people love to be led by novel ideas regardless of their basis in reality.
It's almost like by insisting that there is a subjective experience (as opposed to what?) you close your eyes to how that experience arises, and prefer instead to implicitly attribute it some magical attributes such that it can never be explained mechanically.
Everyone has a different subjective experience of everything because our neurology is different. No person has the exact same arrangement of rods and cones in their eyes, or cilia in ears, or anything like that. No two people have the same arrangement of neurons or dendritic connections in their brain. No two people have the same weights to each of their neurons, so to speak from a ML standpoint. And because of that, all of our experiences are unique and subjective.
BUT, we are SO similar genetically that there is SHARED properties to nearly all experiences among ALL humans, strongly arguing for objectivity to it.
Why is red specifically what we perceive as red is not answerable. How it affects you cognitively and emotionally is based on the entirety of your upbringing and past experiences as well as your sense neurology. Our brain has quite a lot of neurons and connections! Trillions of connections and billions of neurons! A lot can change and be affected in innumerably complex ways! It's more than sufficient to explain the entirety of our subjective experience.
If you are controlling a computer, that control is not visible within the software of computer. The state of computer is accessible to you from "outside", say, by reading hardware state. And you can manipulate that state directly from outside.
This control and access, which needs a true observer, is consciousness. If a computer is taken as layer of reality, then this control is in outside layer of reality. But we don't stop there: It is observer in all layers of reality, whatever reality is made of. The observer can never be represented in any layer of reality.
We should consider that experience layer, like sound and vision are an independent layer of reality than underlying "material" reality. A materialist will say latter exists a priory, and thus experience is off the material layer. But we can as well argue that experience layer is a-priori. My taste sense, for e.g. has no equivalence in material reality. It deserves its own layer of reality. If I hear a voice of someone, it is singular experience rather than experience of collection of frequencies. Color like white is an experience of multiple frequencies of light. Our perceptions, happening with help of neurons, are not based on light at all. Projection is a-priory present, from which we create even light and other formulations...
A mathematical reality is in its own layer of reality. A conceptual reality of say, chair, is also in a "mental" reality.
So we should assume that all are simply abstract layers of reality. But our mental conditioning has created "things" model, just because we can touch and feel, or perceive a boundary in what we see.
As observer, which creates memory of observation, and accesses it, all in the present. An observation which is not in time and space, but creates sense of time and space out of observation.
And so on... A totally different model of reality.
Cognition and consciousness are not easily separable. Some psychology and neuroscience publications that I found particularly good: Gibson (1950), Ittelson (1973), Kosslyn (2005), Northoff (2004) ("Mind–body problem"), Dewey, & Bentley (1949). Particularly the first and last should get you started.
[1]: Gibson, J. J. (1950). The perception of the visual world. Boston: Houghton Mifflin
[2]: Ittelson, W.H. (1973). Environment and Cognition. Seminar Press, New York, NY, 187 pp
[3]: Kosslyn, S. (2005). Mental images and the brain. Cognitive Neuropsychology, 22, 333–347
[4]: Northoff, G. (2004). Philosophy of the brain: The brain problem. Amsterdam: J. Benjamins Pub.
[5]: Dewey, J., & Bentley, A. F. (1949). Knowing and the known. Boston: Beacon Press
"Philosophy of mind" is the discipline you're looking for. They're far off from a preliminary theory let alone a coherent definition of the word consciousness. It is not an easy problem.
A prominent figure in the space is David Chalmers, who made it his life's work to characterize and explore theories of consciousness. He's a philosopher in New York (originally Australia) He coined "The Hard Problem" and "The Meta Problem". You should get an overview in his books. They're a bit dated but it's not like there's been substantial progress.
There's a recent book called "Conscious: A Brief Guide to the Fundamental Mystery of the Mind" which purports to be what you are looking for. She's an author who is married to Sam Harris, a neuroscience PhD and public thinker who focused a lot of consciousness. His ideas permeate her thinking and I found her treatment to be less open minded.
There's a series mirrored on YouTube called "Closer to Truth", where Robert Lawrence Kuhn interviews famous physicists, mathematicians, etc. on this and other grand unsolved questions. You can find interviews with most of the analytical thinkers in the consciousness space if you wade through.
The most "mathematically rigorous" theory is Integrated Information Theory by Giulio Tononi. The logical conclusions of his theory are outlandish or eye-opening, depending on your bias. I don't buy it myself but the there might be something in the approach.
The most mathematically gifted individual with a theory is Roger Penrose. His scientific work is absolutely brilliant but his ideas about consciousness are quite frankly ridiculous and unmotivated.
The best, most rigorous theory I know of comes from a Neuroscientist at Princeton named Michael Graziano. He put forth the "Attention Schema Theory". I strongly recommend his book "Rethinking Consciousness" which makes a case for it, and provides an overview of the other approaches and how they relate to his ideas.
I've been very much swayed by Graziano's work, and we have the technology today to test his theory by integrating actively controlled attention mechanisms into machine learning models. If his ideas are correct, that could lead to models organically making similar claims to humans regarding phenomenological states. The theory is agnostic about whether the claims would be substantiated.
To counter Chalmers, read Dennett and "Consciousness Explained", which is the basis of Graziano's work. Good to see another fan! I posted his latest review paper earlier in this thread.
“Altered States of Consciousness” by Marc Wittmann was an interesting read for me, give the description a read to see if it aligns with what you’re looking for.
Altered States of Consciousness: Experiences Out of Time and Self (The MIT Press) https://a.co/d/ghPhlDW
A somewhat unpopular but strictly pragmatic view on human non-specialness and what that means for consciousness is eliminative materialism, where the qualities we attribute to consciousness are actually illusory, like other sensory illusions. This puts it strictly within the bounds of known physics, no magic.
This paper is a good, accessible overview of how this would work in neuroscience:
I would like to learn how to translate relatively high-level ideas like in the presentation you shared (thanks!) to actual equations to code. Are there any good resources that you would recommend (YouTube channels, books, etc.)? They don't have to be consciousness-related but I would prefer to avoid books on modeling, say, mechanical things because they usually end up over-relying on the reader's background, e.g., in physics. I am more interested in learning how to express "novel" ideas in equations and code.
Consciousness is a quantum phenomenon. The mathematical structure is geometric in some way. The quantum domain acts as a sieve. Constructive and distinctive interference perturbs the quantum substrate, which is the existential universe itself.
Consciousness is thus the universe peering back into itself, and living systems are the biotechnology which animate this phenomena.
The unique properties of this are:
Our “computation” is non-linear. The quantum coherence creates a holographic subjective singularity. These allow the illusion of one perspective among billions of participating neurons.
The more neurons in a cluster, the higher resolution the “pocket of consciousness”. These bubbles compete in the mind for influence. I say this acts much like a hologram.
The sieve of consciousness works through constructive and destructive interference.
Geeze, that’s not how truth works actually (not confined to the horizon of ignorance), though I accept your skepticism.
Your downvotes are your HYPOCRISY!
I have a whole long list of assertions you will fight me on for the love of your ignorance.
In this moment I cannot fully invest in exhaustive debate.
I will prattle on a little more in an unstructured form for the sake of intrigue.
There are several things going, one being this parallel discussion on quantum entanglement (and the parallel discussion for our minds being hackable for another time.)
I would like to describe to you the technology of consciousness in ways I have come to understand, yet I will expect to be scorned and badgered for the inherently abstract nature of our topic.
Neurons in isolation exhibit a telemetric awareness of one and other
The Cosmic Serpent was inspiring in my youth. The primitive ideas the author discusses relate.
Microtubul structures likely provide a substrate for our experiential being. Electrochemical processes flash these substrates, and then gently and delicately suss the state resolve as though a sieve + stack.
The sieve is the constructive /destructive aspect.
I say “stack” as resolve is apparently pushed and pulled or read/write like memories, yet I believe this is where our non-classical thought strategy arises. The outcome is more like a micro-inference rather than a faithful memory.
Constructive and destructive patterns form an analog isolated holographic “image” (though not merely a picture, in every shade and depth of every experience we are capable of having.)
Any information can be pushed into this stack, yet interference collapses to associations (or novelty) producing a primitive psycho-emotional echo, other parts of the brain process into higher level responses.
Consciousness is not intelligence. Consciousness coopts intelligence as a tool for perpetuation.
How does this not sound like consciousness: “consciousness is the inflection upon the potential of our existential being.”
Existential being is any matter, yet all cold matter is dormant without animation through life or other technology.
The potential is the hyper-dimensional capacity of the carefully arranged atoms in our neural substrate.
Inflection is any perturbation of the quantum dynamic (which is usually pure chaos.) it is the atom in the universe which exists, the living system is a technology which extends and allows the aggregate scope of potential awareness and feedback.
The quantum layer is like an echo chamber for your jelly bag brain. This echo chamber is the material universe (which you are) echoing experience upon its analog subjectively scoped potentials.
Whenever I watch a western movie there's always this character type who uses much more complicated words and convoluted sentences than his counterparts. On Tvtropes they call it the "Southern-Fried Genius" and "Sesquipedalian Loquaciousness" (Damn I love this website, another lost hour)
I always love to imagine how fun an exercise it must be for a writer to invent those dialogues.
As an unexpected consequence of your illadvised and unfortunate mentioning of the cyberspace resource which shall not be named for fear of it corrupting the minds of unsuspecting victims and devouring their temporal currency, I have discovered 'The Thing That Goes "Doink"'.
I didn’t really mean to question the phrase hypothesis, rather that “lacking proof is not proof of lacking”, and why not lay out more form of curious extrapolation if that is after all, why we are here.
I DEFINE CONSCIOUSNESS AS THE INFLECTION UPON THE SUBSTRATE OF EXISTENTIAL BEING!
B) I say all life has some smidgen of this magic. It is non-classical and will not emerge out of complexity. All humans do not have “an equal amount”. Next argument may be of capacity.
The universe is the medium, consciousness the inflection, life the technology which animates.
C) voices in our heads. I know, I know, you don’t want to believe that either. Pay attention, check back in ten years. Hear them? Doubt we’re being screwed with? That old “master mind theory” from a hundred years ago touched upon this before our taboo of speaking of these voices dismissed them as crazy. Snail telegraph? There are clues throughout history.
We’re entangled, and it’s more of a hassle than a gift.
Definitions are not magical bindings that would guarantee the adequacy of a representation but the fixed points assumed to aid communication, science is not a 'search' for definitions, it uses definitions to communicate what it has discovered.
The mathematical structure of phenomenology is kind of a weird specific task, because it's traditionally equated to the structure of the world or representation in general. This is because in theory 'consciousness' means very little when abstracted from the theory of its implementation, on the other hand if we try to preserve the relation of consciousness with the implementation we already have to choose between several. I think the best candidate for the existence of some of most rudimentary sentience is the gamma oscillations that function as coordinator functions of different neural populations, so the math to represent this would be take that as a basis.
The fundamental flaw with this entire branch is the premise that consciousness is mathematical. Math is a tool of reason, not the cause. And once you adopt this premise, it makes all progress toward understanding reason impossible especially how concepts work, the key differentiator between humans and other animals without conceptual faculties.
> term ‘aspect’ denotes an instantiated phenomenal property or quale. The set of aspects A(e),
> then, comprises the phenomenal properties or qualia which are instantiated in the experience e,
> also called the phenomenal states of the experience e. Our question, then, is what it means
> that “any set of phenomenal states of a subject at a time is phenomenally unified”
> [...]
> A promising answer is the
> so-called subsumptive unity thesis, developed in [5]:
> “For any set of phenomenal states of a subject at a time, the subject has a
> phenomenal state that subsumes each of the states in that set.” [5, p. 20]
> According to this thesis, what underlies the experience of the unity of a conscious experience
> is that for any set X of phenomenal states in the conscious experience, there is a further
> phenomenal state that subsumes each of the states in X. This phenomenal state characterizes
> what it is like to be in all of the states of X at once [5, p. 20].
> [...]
> Phenomenal unity gives rise to a mathematical structure of conscious experience.
When I try to state this in natural language I come to something like "Conscious states consist of qualities, or contain qualia. And they are unified _in_ consciousness, meaning that our experience is whole. The mathematical structure is this: For any set of experiencing qualities there is a(n additional) state integrating them all into one state."
When I studied philosophy in the 2000s we discussed such kinds of solutions already in a seminar and found them lacking on two grounds.
First: there is no clear notion of non-conscious qualia, and Ockham help us here to keep it so. But if this is true, each of the integrated states already presupposes the whole (because it is already conscious, consciousness is more the medium of qualia), which then cannot come after them to unify them. Unless one wants to assume that a lot of conscious splintered experiences are integrated into a whole later. That would be even more questionable, because who experiences them?
Second: The integrating state needs another integrating state to integrate it into the whole of conscious experience etc. ad inf. Unless it is a state that integrates itself into the whole as well, which would just be a sleigh of hand - because exactly that is the riddle that a mathematical structure must overcome to be about consciousness. The self-reference, the self-presentation of consciousness. Apologies for lacking the proper terminology here.
Yeah, I think that for one conscious experience e they would like to say not only are there a bunch of qualia involved (the "aspects" of e: A(e)) but that, in Nagel's famous phrase, there is something it is like to be the subject experiencing A(e) in the same instant, over and above the set of qualia A(e) itself.
I agree that "non-conscious qualia" is an incoherent notion, but I'm not convinced this rules out a "phenomenal unity" as you argue - they are positing the unity as "phenomenal" not as the substrate (or medium, as you put it) of qualia, so more than just the raw fact of consciousness is implied.
One point of comparison might be Wittgenstein's famous example of the duck-rabbit - a bistable ambiguous image that can be seen either as a picture of a duck or as a picture of a rabbit. One might assume the qualia associated with looking at the picture are the same whether one is seeing it as a duck or as a rabbit - the same shapes and colours - and yet there is a difference:
> The change of aspect. "But surely you would say the picture is altogether different now!" [...]
> If I saw the duck-rabbit as a rabbit, then I saw: these shapes and colours (I give them in detail) -- and I saw besides something like this: and here I point to a number of different pictures of rabbits [...]
> 'Seeing as ...' is not part of perception. And for that reason it is like seeing and again not like.
Perhaps this un-perceptual yet pervasive and arguably phenomenal unity subsuming the duck-rabbit qualia - that which flips from duck to rabbit and back - is analogous to what the authors intend with their "subsumptive unity thesis".
Congrats on commenting on TFA, btw :) I thought it was an interesting read, though I start to lose my grip on the math around the topology section.
Consider TDVP (Triadic Distinction Vortical Paradigm) from Ed Close (erclosetphysics.com) and Vernon Neppe (vernonneppe.org), based on the work of George Spencer Brown, who invented the Calculus of Indications which encapsulates boolean logic but does so in a profoundly elegant visual grammar. One that dramatically shortens calculation visually.
Distinction: to distinguish anything from anything else.
Consciousness: the distinction of self from other. Also has further distinctions of inside the self and the other organized into patterns, and distinctions drawn for negative entropy and spiritual evolution, the basis of all change.
Distinctions have three parts, that which is distinguished, that from which it is distinguished, and the consciousness drawing the distinction. This has reality differntiation along the lines of the Perceptual, Conceptual and Existential, and distinctions themselves have variables of Intent, Content and Extent.
Imagine a idealist, unified monism 9-D model of reality that has three main prongs. Other unstable particles teased into existence temporarily by LHC are not very useful whereas normalization based on the characteristics of the free electron when it comes to Space/Time/Consciousness are relevant.
For symbolic representation of Extent: this is Space that has three dimensions and is represented using integers. And then there is Time, which are image events of cause and effect on a timeline, that are represented like a plane, a volume, or that of more other people, and this represented with imaginary numbers. And then there is Consciousness-awareness, complex numbers, the awareness of being, or the distinction of Self from Other as spiritual awareness or life trace, and it too is a plane, a volume, and that of more other people.
For the symmetric resonance of Content we have Mass (subatomic, atomic and the universe), Energy (photonic, electric, magnetic) and also Consciousness-meaningfulness (individual, group and cosmic) aka gimmel, the third form of reality which is not mass or energy.
Primary Consciousness is one thing, but then gimmel, the organizer, is second and mass/energy is third.
For Intent or meaningful information: existential, conceptual and imaginary.
In this model we are all moving towards experiences like what is termed Cosmic Consciousness (omnipresence) in Yoga, especially along the lines of Kriya Yoga as brought forward into the West by Yogananda. Anyway, perhaps a start towards mathematizing consciousness that is compatible with ancient wisdom and modern mathematics.
178 comments
[ 2.7 ms ] story [ 256 ms ] threadNothing is answered in this article. We simply don't know.
This article is up here because nobody is sure whether or not chatGPT is in any way representative of an aspect of our conscious experience.
Is chatGPT a mathematical model of consciousness?
We know it doesn't represent the whole experience because... it has huge flaws... but we aren't sure whether it represents an aspect of what we are.
Disagree or agree, we have something in our hands and it could be it... but at this point we don't know.
To be clear, we can't yet conclude that these huge flaws are due to its structure rather than the extra sensory data that we have and it doesn't. Doesn't really change your overall point, just wanted to clarify what we can and cannot conclude from the evidence, because people are far too quick to jump to conclusions after seeing the flaws.
In any case, it is still just a lookup table. Compression, unless lossy, doesn't change the meaning of the information in this lookup table. It just makes access more indirect.
Hutter might think that intelligence is a form of compression, but I think that's just wishful thinking. It might be my ego speaking to say I think I'm the one that's right. But it's not my ego speaking to say that there's no hard proof to say unequivocally that intelligence is equivalent to compression.
It doesn't have to be lossless compression. Intelligence is also about being discerning about the information that's relevant, although that's basically the same point.
> But it's not my ego speaking to say that there's no hard proof to say unequivocally that intelligence is equivalent to compression.
Actually I think it's kind of trivial to see that intelligence is a necessarily form of compression. Abstractions, concepts, theories, these are all lossy descriptions that discards irrelevant details and capture only relevant details to make problem solving via deduction, induction or abduction tractable. And yet, using those concepts, theories, abstractions you can reproduce/predict observations with very high fidelity. That's compression.
Do you have a counterexample of something that we consider intelligence that doesn't fit this model? I suppose it depends what specifically you mean by "equivalent".
It is unclear what necessitates any closed-form mathematical description of 'consciousness'.
Such a concept seems tantamount to a faith claim.
No, we haven't clearly defined consciousness. For a given exact definition of consciousness, we can determine whether something fits that definition or not. If we don't clearly define it, then we're just talking about a feeling, a qualia, and it's not possible for people to reach a consensus because they're not even discussing the exact same thing when they say "consciousness".
We agree that humans are conscious, we agree that rocks are not conscious. This classification is clear and points to a underlying definition we cannot fully articulate. It is more than a qualia or whatever you call it. We know it has exact parameters simply because we are able to perform classification tasks.
The consensus breaks for things like chatGPT. chatGPT straddles some uncertain border that we can't articulate. But in general for the term consciousness we reach consensus by how we classify things as conscious or unconscious.
Clearly there's not agreement in what is classified as conscious, given this thread/article.
>We agree that humans are conscious, we agree that rocks are not conscious. This classification is clear and points to a underlying definition we cannot fully articulate. It is more than a qualia or whatever you call it.
This is a fallacy. People agree that yellow is not violet and green is not violet, but they don't agree that indigo is not violet. Just because there are some classifications on which people agree about the label this doesn't imply they agree upon all classifications. A more direct example is whether plants are conscious or not.
There is general agreement. We agree for most things like rocks and humans..there are a few things where we aren't sure, like chatGPT.
But in general humans are in agreement for most things.
>Just because there are some classifications on which people agree about the label this doesn't imply they agree upon all classifications.
Well just because ambiguities for certain things exist doesn't preclude the existence of things that are utterly clear to be conscious or unconscious.
We know humans are conscious and we know humans have all the required parameters to make them classified as conscious. We just can't deduce the exact thresholds and features that cause a phase shift when changed. But we know these parameters exist.
Unless we reduce models of consciousness to purely stochastic processes, this is trivially false. And no one seriously reduces consciousness to probabilistic pattern recognition.
> We know it doesn't represent the whole experience...
It doesn't represent any experience. ChatGPT doesn't "experience" things as conscious things do. It's not even comparable to the experience of a simple invertebrate, for example.
I think consciousness has to be the product of an algorithmic process. I think this process can be summed up as "prediction". I don't think there's some secret quantum/soul/magic aspect to it. Which is why I think it's plausible that transformer models may capture some aspect of what we refer to as consciousness, though I don't quite buy it at a gut level.
You say that "it's not even comparable to the experience of a simple invertebrate," but that's just you writing down what your gut is telling you. What is the experience of a simple invertebrate like, exactly?
The question rattling around in my brain tonight is - if some, many, people don't develop a theory of mind as young children, how many of the people around me are like this? Is it just the philosophers?
It seems of much more moment than the essence of ChatGPT.
Given that, I'm not sure how to interpret your comment as it seems to describe an invalid inference.
Most things in this universe have mathematical models that can describe it. We have models for virtually everything. Thus it is reasonable to say such a model also exists for consciousness. To say what you said definitively without even knowing the meaning of the word consciousness is premature. You don't have enough knowledge to say anything about this.
>It doesn't represent any experience. ChatGPT doesn't "experience" things as conscious things do. It's not even comparable to the experience of a simple invertebrate, for example.
Again you don't know what you're talking about. Nobody can define what consciousness is and nobody fully understands the emergent effects of chatGPT or the human brain. To say what you say with conviction is therefore empty. You don't actually know because nobody knows.
I think that's a bit uncharitable. You're acting like people haven't been grappling with these problems for thousands of years. Searle, for example, makes a pretty good argument why symbol-processing systems will never produce consciousness (you'd have to prove that the Chinese room is conscious, etc.).
It's clear, for example, that a toaster doesn't have any conscious experiences, but a dog does. To say that "nobody fully understands" the human brain is throwing out the baby with the bathwater. ChatGPT is not conscious any more than my (very complex) Swiss wristwatch is.
If an atom can be described by a mathematical equation that means so can groups of atoms. If groups of atoms can be described math and if a neuron is simply a complex group of atoms then a neuron can be described by an equation. If neurons can be described by an equation then so can groups of neurons. If the brain is a group of neurons then the brain can be described by a mathematical equation. If consciousness is a physical phenomena produced by a mathematical brain then consciousness is a mathematical model.
The only way you can believe that consciousness cannot ever by described by math is if you see some break in the previous chain of logic. Most likely it's in the linking of consciousness entirely with the physical brain. Likely you believe in a soul or something if you don't agree with the logic. Which just means you're using mythology and concepts outside of science to derive your conclusions. Again, lighting is not thrown down from the sky by a god named Zeus.
It's clear a toaster is not conscious. But what chatGPT is, is not clear. Placing it in a category where we aren't sure is reasonable and also it's a category agreed upon by many people. It would be a bit uncharitable to demand that your categorization of chatGPT is more correct then other peoples. Even you must be be aware of the ambiguity associated with chatGPT despite your stubborn stance to the contrary.
I would posit that it is clear. It's simply a whole bunch of toasters put together (that is, a long list of if-then statements). If you're trying to argue that consciousness is merely a bunch of if-then statements, then yes, ChatGPT might be conscious. But then it's going to be much harder to argue that a toaster isn't conscious, which I thought we both agreed with.
> Likely you believe in a soul or something if you don't agree with the logic.
This seems a bit dismissive, and although I myself am a substance dualist (which is generally seen as a metaphysical position), most materialists aren't mechanistic in their interpretation of consciousness.
The combination of multitudes of if-else statements forms a macro structure. It is this emergent macro structure that we call consciousness. As stated before the exact interconnect of this macro structure can't be fully articulated yet.
>This seems a bit dismissive, and although I myself am a substance dualist (which is generally seen as a metaphysical position), most materialists aren't mechanistic in their interpretation of consciousness
My arguments are based on likelihood derived from logic and evidence meaning While we don't fully know if a soul fits in this framework, we certainly do not yet have the capability to derive it's existence/likelihood from evidence or first principles. Thus there's really nothing to talk about as it's an argument outside of known logic. What is the point of arguing about topics we don't know about? There is no point. I'm being dismissive because there is literally nothing left to talk about for either of us. Logic and evidence is also the only shared principles that seem consistent across individuals so best keep discussions in that realm fully.
It can be, and it's called the theory of computation. It's been extensively studied for the past century and a half, and, all-in-all, it's not even particularly strong. For example, most real numbers aren't even computable, Godel showed that consistent theories can't even prove their own consistency, and so on. In my view, our mind does not seem to compute things (though some, like Dennett and Putnam, would disagree).
The claim that our mind computes things is a pretty strong one (and if correct, I think we'd all probably be much better at math). Penrose argues that whatever the mind does, it's (at least in part) a fundamentally unsound operation; this to me sounds way more correct than the idea that our mind operates algorithmically.
It's the most likely outcome given what we know of the physical world.
> Godel showed that consistent theories can't even prove their own consistency, and so on.
Godel showed a flaw within logic itself. But it's all we got. If you're operating outside of logic then we don't have a consistent logical framework to use for generating conclusions. Without this, like I said, the conversation is pretty much over. There is nothing left to say.
Even so. For us, consciousness is just there.
Externally we attribute consciousness to certain behaviours, including (but not limited to) self-reference, goal-seeking, physical and social responsiveness, use of language, memory, and so on.
These two things are not the same. We assume other humans are conscious in the same way we are to the extent that they perform consciousness. We can't be sure, but for human interactions it's a workable proxy.
A toaster, a watch, a rock, and a web server don't perform consciousness in any way at all. They may have some kind of metaphyiscal awareness, or not, but if they do we can't see it, and without a comprehensive theory of consciousness it's parsimonious to assume they're not aware.
ChatGPT performs some elements of consciousness, but not others. It happens to perform the elements AI researchers consider important - specifically use of language. But it doesn't have a memory for previous conversations, it doesn't show any independent goal seeking, and so on.
It uses "I..." and a vast amount of human training to mimic the appearance of consciousness, and plenty of humans seem to want to believe that's enough.
But the essence of the hard problem is the difference between "performs the actions we associate with consciousness" and "is conscious in the same way we know we are."
You could argue that practically there's no difference, and from one point of view that may be true. But that doesn't work if you want to understand what's happening rather than just assuming in a "well obviously..." kind of a way.
If only because historically in science "well obviously..." has been consistently wrong.
Another option is there is some kind of metaphyiscal awareness everywhere, and it's only coincidental that objects of class "human" eventually convince themselves into believing that it's equivalent or highly related to their own thinking process. Unless we exclude this option somehow, AI can't be definitely seen as simply rolex, because it does that too.
Currently AI is discontinuous and passive, but is that even required?
without a comprehensive theory of consciousness it's parsimonious to assume they're not aware
This hypothesis creates more entities (one island of awareness per complex enough object), so why not the other way round?
So anthropic. We can agree that it is on the opposite end of a spectrum though. A spectrum that we define out of a single point (“us”, and mostly “I”). We can also say that a toaster is a part of your consciousness instead, like e.g. your arm but less noisy. But then you are a part of a dog’s consciousness as well.
Consciousness is probably best described as just a field in a complexity space, not a discrete set of localized phenomenons. It has a near-zero value in a physical projection of a toaster’s concept, but is that projection even useful?
Why is it that optimizing for survival can lead to consciousness, but not predictive coding? Nothing here is trivial.
Moreover, large language models (LLMs), like GPT-3, are designed to learn how to predict, and prediction is the dual of compression. As explained in this StackExchange post (https://cs.stackexchange.com/questions/130655/is-prediction-...), prediction can be used to compress data, and compression can be used to make predictions. Given that LLMs are trained on vast amounts of data and can predict patterns within that data, I believe that they contain all the necessary components to be considered an intelligent system.
Therefore, I firmly believe that consciousness can indeed be reduced to probabilistic pattern recognition, and that LLMs provide a compelling example of this phenomenon in action.
Computation is not necessarily experience. I can't emphasize enough how fundamental understanding this is. It seems some just can't grasp this concept.
Would you say a program that echoes to the screen the responses of a hidden human is intelligent? Viewed as a black box, it would certainly seem so. But does it have any internal experience and understanding of anything? Of course this is a trivial experiment, but the point of it is to show, in the most absurdly clear terms, that producing plausible outputs !== being intelligent, much less having a conscious experience.
This example also shows that the perceived sophistication of the input-output transformation is immaterial, unless you want to say consciousness requires a certain internal architecture for computation, in which case I would ask: what architecture is that? No matter the sophistication of an algorithm, it's computation lies on a different axis to that of consciousness. You might say human compute because they are conscious, not conscious because they compute.
Another way to understand this: Let's say you touch a very hot object. You skin, in conjunction with your brain, senses and computes the temperature of the object. But, there is no essential reason why that should lead to an experience of pain, rather than a sensation similar to color, which generally does not produce pain.
No, it doesn't show that, it merely asserts it. Consciousness could be a logical property of some perceptual algorithm, like ordered outputs is a property of a sorting algorithm.
> But, there is no essential reason why that should lead to an experience of pain, rather than a sensation similar to color, which generally does not produce pain.
Sure there is, because it's approaching the point of causing damage. Pain happens when sensory precepts approach acceptable tolerance. Pain is like clipping in an audio stream.
I’ve given you an example that proves it.
> Sure there is, because it's approaching the point of causing damage.
There are plenty of damaging processes in the body that produce no pain.
But that’s not even the issue. What I’m interested in is how sensations are produced; it’s not just computation
Except you didn't. The black box you described does have feelings because the human inside has feelings. The fact that every component of a sentient system doesn't have feelings isn't particularly interesting.
> There are plenty of damaging processes in the body that produce no pain.
Because they didn't need to for fitness purposes.
> What I’m interested in is how sensations are produced;
https://www.pnas.org/doi/10.1073/pnas.2116933119
> it’s not just computation
You don't know that.
Huh, damaging process are not relevant for fitness purposes? Wow
Your link is full of speculation, and indeed for consciousness, speculation may be all that can be done, since no one can define it, only experience it.
> You don't know that.
It seems self-evident. If a program computes the temperature of the sun, does it feel the heat?
Cancer or heart disease in your 50s does not impact your reproductive fitness, for instance. Way to attach a strawman.
> It seems self-evident. If a program computes the temperature of the sun, does it feel the heat?
The sun isn't computing a temperature, it's computing a nuclear fusion reaction. Also, "feeling" isn't a quality of individual particles to the best of our knowledge, but it is a quality of some aggregate of particles we call a human.
So if your question is whether a faithful simulation of a human would feel the heat from a faithful simulation of the sun, then yes, that's very plausible.
But there are plenty of cancers that can and do occur in ones reproductive years, that start without pain.
> The sun isn't computing a temperature, it's computing a nuclear fusion reaction.
I wasn’t comparing the program to the sun. I was comparing it to the human skin.
> Also, "feeling" isn't a quality of individual particles to the best of our knowledge, but it is a quality of some aggregate of particles we call a human.
Hmmm, now you are referring to particles and aggregates. Wonder what the source of that claim is. At least you agree computation (a process) is not the important thing, but particles (matter).
> So if your question is whether a faithful simulation of a human would feel the heat from a faithful simulation of the sun, then yes, that's very plausible.
Wow, now you're making a circular argument. Whats a “faithful” simulation of a human? How would we know it is “faithful”? What form would that even take?
Let’s say ChatGPT becomes “sufficiently” advanced, and is hooked up to a sensor that records the temperature of the sun. Would it feel the heat? The whole idea is preposterous.
So? The incidence of these cancers is insufficient to impact reproductive fitness in our species. At no point did I claim that all damaging processes must produce pain, I only claimed that all pain originates from evolutionary pressures indicating high chance of damage that does impact fitness. That also does not mean that presence of pain is always directly indicative of damage, because biology is messy and precision has a cost to fitness.
> At least you agree computation (a process) is not the important thing, but particles (matter).
Nope, not what I'm saying at all. The particles are a computation in a simulation, so the information content of the computation is what matters.
> Wow, now you're making a circular argument. Whats a “faithful” simulation of a human? How would we know it is “faithful”? What form would that even take?
A faithful simulation is clearly a simulation that contains all of the information content of a human and replicates the mathematical relationships that govern the operation of a human's mind. Since a human is a bounded volume, it contains finite information per the Bekenstein Bound, and is thus amenable to simulation, unless you are asserting some kind of non-physical dualism.
It seems clear that you don't yet grasp the argument being made so maybe you should reserve judgment, like calling the argument "circular", until you do properly grasp it.
> Let’s say ChatGPT becomes “sufficiently” advanced, and is hooked up to a sensor that records the temperature of the sun. Would it feel the heat?
It wouldn't feel heat in the same way that humans feel heat because its sensor and information processing is not mathematically equivalent to human information processing and sensory systems. That does not mean it wouldn't feel something like its own version of heat. A whole lot of vagueness hides behind your "sufficiently advanced".
If ChatGPT v10.0 did reproduce the mathematical relationships underlying human information processing and sensory systems, then yes, it would feel heat in precisely the same way as humans do.
Think about it, if we accept the theory that computation equals consciousness, then then is little to investigate further. Just calculate something, and voila, consciousness exists.
It is this closing down of inquiry that I push against. My assertion is that consciousness is deeper, it cannot be reduced to computation.
How would you even prove otherwise? If it arises from computation, what is conscious? Is it the algorithm? Is it the electric circuits? It is the physical machine?
It is interesting to note my consciousness is bound up what a notion of what my body is. I can identify the limits of my body and make a distinction between it and my environment. What determines the conscious boundary of a entity performing calculations?
On the other hand, there are many strong arguments against consciousness arising from mere computation, some of which I have enumerated.
All eliminative materialists disagree. To them, consciousness is mechanistic in some way, so why not stochastic pattern matching which can be made Turing complete quite easily. For instance: https://www.pnas.org/doi/10.1073/pnas.2116933119
> It doesn't represent any experience. ChatGPT doesn't "experience" things as conscious things do.
You don't know that because we don't have a mechanistic understanding of "experience".
IMO the hard problem is not posed to be solved. I think of it like this. The world in which we are p-zombies is categorically indistinguishable from the one in which we are not, by an objective perspective. It is not a concept that we should use for an objective definition.
It's like saying "a computer that works, but where the CPU is replaced by a piece of stone is physically indistinguishable from the real thing".
We have not yet found a physical distinction. It seems likely that there is one.
https://www.engraved.blog/building-a-virtual-machine-inside/
Read to the ending where it closes the recursive loop.
I'm sure Hofstader would dislike people saying "he was right about this!" "he defined this!", though, since that's not what he was setting out to do at all. GEB really is just an exploration, with some of the results being couched in mathematical language as an artifact of his background, not as an attempt to formalize anything.
You want an explanation for the sensation of the color red? Alright, I'll try, as long as you help me out: what exactly is the sensation of the color red? I need to have a good grasp on what we're talking about here.
Can you describe it? It's okay, take your time.
FIN
And of course the entire point is that qualia are ineffable. That's part of the definition. It's the feature, not a bug. But if you want to deny that you have conscious sensation then go ahead. ;)
Anil Seth's Being You is a recent one that does a good job of summarizing the most popular theories in consciousness. https://www.goodreads.com/book/show/53036979-being-you
The Routledge Handbook of Consciousness (https://www.amazon.com/Routledge-Handbook-Consciousness-Hand...) is far more comprehensive and while being more academicky is still an excellent read.
Unfortunately both miss what I think is easily the best and most computationally robust and fleshed out consciousness theory, but that's because it's so tragically unknown. I co-authored a book that covers this (while staying accessible) https://www.goodreads.com/book/show/58085266-journey-of-the-...
So much of the stuff in the mathematical-this and physics-that of conscious experience is just crap because there is very little understanding of biological consciousness and what it is and where we can find it, and where we cannot. These start with assumptions of what consciousness is and then go on to mathematize these non-validated assumptions. (Integrated Information Theory) IIT is the most popular among these. https://scottaaronson.blog/?p=1799 -- this article captures well why IIT, and many like it, are bad models of consciousness.
I am a software engineer by profession, but this is a topic that has been on the back of my mind for many years. Even when I read popsci books about the universe and the big bang and so on, i felt spiritually unsatisfied because they often do not address this elephant in the room - consciousness. A theory of everything that explains the laws of universe will be imo incomplete if it doesn't also explain, at the fundamental level, the laws of the universe that exists behind our eyes. It really deserves more attention and awareness among geeks and the general public.
But I agree with your assertion.
Galileo’s Error, published in 2019, by Phillip Goff traces the history of the mind-body problem and illustrates how we divorced our consciousness from the physical world during the development of early modern science and philosophy.
Last century, Ernst Mach and Bertrand Russell argued in favour for neutral monism, a stance which holds that the mental and physical are two sides of the same coin. It seems like these ideas are coming back after a long stretch of dogmatic physicalism.
Panpsychism gets a bad rep for sounding like it has associations with vacuous or flimsy New Age beliefs, but it means a totally different thing in the context of philosophy of mind and consciousness studies. It is the claim that consciousness might be a building block of the universe.
The main problem with panpsychism is not that it is incredulous, but that it currently lacks a compelling explanation for our individual subjectivity. If consciousness and mental properties are ubiquitous in the universe, why do I have my consciousness and you have yours? The main motivation to believe in panpsychism seems to be its incredible elegance and self-consistency. That alone does not make it true, but it might be a very big hint on where to look for answers.
You are just assuming that LLMs aren't conscious. We don't have a way to evaluate that at this time.
> It really deserves more attention and awareness among geeks and the general public.
There's 0 interest in consciousness because it's so hard to pay attention and realize that it exists. We go about our days for weeks, months, years and even decades without even noticing that consciousness is a thing.
Do we? I feel like I'm constantly reflecting on it and trying to make sense of it, and that I'm certainly not alone. I'd like to think that it's part of being human to turn the lens back on itself and wonder "what am I?" But I suppose some people could care less for navel gazing.
When really, humans give themselves (their own thoughts) way too much credit and most are barely aware of their own thoughts, barely self-aware at all.
It is easy to say something is ‘un-explainable’ and thus ‘must be in some otherworldly realm that is beyond comprehension’, if one is not able to observer the problem. Most humans aren’t self-aware enough to know their own thoughts or examine where those thoughts come from.
Just a few decades ago simple voice recognition was deemed ‘a totally human ability that the computer will never accomplish’, now it is ordinary.
It comes down to humans internal subjective reality is both amazing and ‘unknowable to the outside’, ’un-observable’, thus un-testable. But then we make this giant leap that nothing we will ever build will also have some ‘internal subjective reality’.
It keeps coming up that GPT is a glorified auto-complete, but then never explain how a human isn’t also just an ‘auto-complete’ machine, just going through the day twitching at stimuli. Why do we think a human doing copy/paste, and regurgitating something they heard, is special, but if a computer does then suddenly it is completely invalid to think they could ever be conscious.
However, OP was referring to the hard problem of consciousness. The hard problem of consciousness is unsolvable IMHO by science, as it is entirely subjective; in fact, it is the fact of existence of subjective experience. The problem is this: even if you find some objective mechanism that generates consciousness, how would you make the explanatory leap from the objective mechanism to the subjective? Even if you say: "Oh, we flipped this switch and now this person is unconscious." - this is a description from the outside. You cannot claim by this that the person's consciousness has disappeared. Maybe not connected to the contents of the human mind anymore; but the gulf between objective and subjective can never be crossed by objective science.
If I had a dollar for every time someone claimed that the hard problem of consciousness isn't actually hard, I would be very rich.
There is only 1 consciousness. Soft is describing how it can be done physically, how to build it, or explain it, and hard is what it is like to experience it from the inside.
By creating the hard problem, it really just says "we can't know anything about any internal state". All "internal subject states" are self reported, and a machine can just as well self report enough to mimic a human. So at that point, we can't prove that the machine doesn't have a 'subjective' internal state.
By this line of reasoning, we can't prove that any humans are conscious either. Any lines of logic that try to prove the hard problem, would either both prove that the machine and human are both conscious, or prove that neither are conscious.
The hard problem is not stating "we can't know anything about any internal state", it is about the mere subjective existence of the internal state and not about the contents of internal experience or any objective fact about that internal state. I don't know why this is seems often so hard, but in consciousness research and meditators' language (or in my native language at least), most of this seems pretty well-definded.
It has no relation to proving that humans or machines are conscious. It is about no objective fact at all, and thus cannot be investigated using objective methods. That is why it is unsolvable by objective science.
To use a koan-like: It is about this.
I would say that this examination of the problem of our internal view of the world, versus the reality of the world, subject versus object, goes back thousands of years.
I've also meditated a long time, and I do understand the hard problem. I just don't think eastern philosophies have the answer either. Eastern philosophies definitely had a jump start on the West with examining the mind, and really it was Buddhism that informed western philosophies from Schopenhauer onward. They helped defining some vocabulary and hierarchies, like outlining Skanda's or parts.
BUT, the typical eastern answer, that it is just 'this" and other 'wooo-wooo' answers are just as far from any explanation to the 'hard' problem as anybody in the west has. Meditation can get you awareness of the hard problem, some insight, but it still isn't an explanation. It just gets you another perspective that is still "Internal". I can meditate forever, but any 'realization' is still internal to me, it isn't providing any answers or explanation.
This comes up a lot right now with AI getting so human like (recent advancements), and I would say that at some point we'll create something that will be indistinguishable from a human, and at that time if the AI speaks convincingly of their realization of 'this is it', we wont have any ability to disprove that they really have had the same 'wooo-wooo' feelings. There is no way to prove anything.
So what I am speculating here, is that at some time in near future, we will be able to correlate soft and hard problems, so we have a map from one to another. And we'll have to realize that our own internal 'experience of reality' is just our neural net in our brain reacting and processing to stimuli, and we'll be able to examine the same things in silicon. And we'll have to accept that we aren't some super mystical beings or that our 'consciousness' is not all that special a development in the universe.
So lets say in current models', or future, in very large neural nets, that there will not be some internal subjective awareness, just like we have. We've mapped the entire neural network/nervous system, of a nematode worm and translated it into software. The 'software worm' appears to behaves the same as the real worm, so what is that software experiencing?
Actually, there is every indication that you are an illusion.
The fact that we think we can observe and comprehend a mechanism doing the observation does not preclude things we cannot see. You might be able to pan a cctv camera and get some insight about the camera itself if you find a reflection, but you can’t pan it over to the linux driver. It is not a given that all of the observer exists within the observable universe.
It seems short sighted to suggest we can’t find out things about consciousness in a similar way.
We can almost definitely learn a lot more about consciousness related phenomenon than we already know through science, and I don’t know the answer to OP’s question about the current cutting edge of that, but I think it’s always going to be about related phenomenon rather than consciousness itself. Unlike other phenomenon, which are only revealed to us because aspects of their existence appear in our observable universe, we don’t really “observe” the core experience of consciousness. The goalposts can be moved indefinitely whenever consciousness related phenomenon are made sense of. The fundamental experience is not observable, which means we can’t really see that black box at all. To me the mystery of that is much more satisfying than thinking we can ever get a definitive answer.
It is easy (and cognitively useful) to forget that, when the indirection is very reliable.
As far as consciousness, we won’t just have to speculate based on our own consciousness. As artificial minds become highly self-aware (a functional ingredient of consciousness), we will be able to construct very controlled experiments.
There will be endless numbers of experiments we will be able to do once we can edit self-aware entities and both externally query and internally monitor them.
That's what the observability issue is about. Heck, I think you're conscious, but I cannot know for sure.
There is no way to distinguish between an unconscious machine mimicking the functional aspects of consciousness we are able to indirectly observe and compare to our own experience and something having the same experience.
We could easily (and I think almost definitely are given the current pure symbolic manipulation track) be doing the same thing as a jungle cargo cult constructing an air strip out of sticks. Studying that won’t tell us how planes work.
Obviously advanced AI is more sophisticated than that and may have more legitimate parallels to consciousness, but I think it’s fundamentally impossible to know what’s a core part of the conscious experience and what just looks superficially similar to the output of conscious beings.
The current generation of AI are clearly not at all conscious and are “just” (don’t want to diminish the work) very good parrots generating image and text output from a given input in a way a whole bunch of conscious humans already have. But we’re already getting cargo cults and people who are anthropomorphizing and considering them comparable to humans.
There’s a danger in studying artificial models of an underlying phenomenon to learn about the phenomenon. That method of study is inherently very limited and extremely difficult to keep calibrated to the underlying phenomenon.
As noted before, all observations are indirect. That has not been the unique challenge for studying consciousness.
The unique challenge has been an inability to create constructive or destructive experiments.
We didn't know how to construct consciousness candidates or precursors, of varying kinds to study. We are getting close to that (i.e. self-awareness as a functional quality is certainly a precursor.)
And we have not been able to create controlled demolition of an consciousness candidates, due to the morals of human experiments. (But our artificial minds will be alterable and destructible at will, to test hypothesis.)
Our conscious experience is in fact a unique phenomenon because it is all perceivable phenomenon.
You can create as many self reflective machines as you like and experiment on them and compare them to output from humans, and maybe you’ll learn things, but you can’t ever get to looking at consciousness at all. There is no indirect or direct data you can use to compare all of your experience to something that sits inside a fraction of it.
What you’re talking about are behavioral experiments comparing the reported thoughts and actions of separate acknowledged conscious actors, ie other people, to the internal states and outputs of machines.
You’re not looking at consciousness when you do that. Looking at consciousness requires looking at all experience. Which is fundamentally impossible. When you try to look at all of experience as a single object to study nothing is differentiated. There is zero information.
The problem with consciousness is that we have no ability to measure it objectively because it is an entirely subjective phenomenon.
You have no way of knowing if I am conscious or just a sufficiently elaborate machine doing a really great job of acting like I am (a p-zombie).
Personally, I lean towards panpsychism. I assume I am not special and that everything in the universe, or to put it another way the entire universe itself, has an experience of being.
Of course not! Regardless of whether the core phenomenon is fundamentally impossible to explain (which I think it has to be for philosophical reasons), it’s an incredibly productive question to ask/am also curious about cutting edge theories and related findings.
(1) https://plato.stanford.edu/entries/kant-reason/#LimiReas
In fact, you would not even experience the disappearance of consciousness - because the consciousness that would experience this has disappeared. It is impossible to experience the absence of consciousness.
https://www.amazon.com/Consciousness-Emily-Troscianko-Susan-...
She has another book called Conversations on Consciousness that's basically transcribed interviews with prominent figures that's also worth checking out.
https://www.goodreads.com/book/show/254499.Conversations_on_...
Yes this comment is an ad hominem, and no I have not read the book... All I have is a skeptical hunch that this might not represent "the state of the art" on the topic.
From her wiki page:
It was just over thirty years ago that I had the dramatic out-of-body experience that convinced me of the reality of psychic phenomena and launched me on a crusade to show those closed-minded scientists that consciousness could reach beyond the body and that death was not the end. Just a few years of careful experiments changed all that. I found no psychic phenomena—only wishful thinking, self-deception, experimental error and, occasionally, fraud. I became a sceptic.
It is also a natural consequence of the culture you happen to have been raised in, which is an extremely interesting aspect of consciousness: how easily it can be programmed, and how influential it can be (it changes how reality "is"). In your case you have self-awareness of the phenomenon, at least in this case, but most people don't operate on that level.
* After spending time in research on parapsychology and the paranormal,[5] her attitude towards the field moved from belief to scepticism.[6][7] In 1987, Blackmore wrote that she had an out-of-body experience shortly after she began running the Oxford University Society for Psychical Research (OUSPR):[8][9]
"Within a few weeks I had not only learned a lot about the occult and the paranormal, but I had an experience that was to have a lasting effect on me—an out-of-body experience (OBE). It happened while I was wide awake, sitting talking to friends. It lasted about three hours and included everything from a typical "astral projection," complete with a silver cord and duplicate body, to free-floating flying, and finally to a mystical experience. It was clear to me that the doctrine of astral projection, with its astral bodies floating about on astral planes, was intellectually unsatisfactory. But to dismiss the experience as "just imagination" would be impossible without being dishonest about how it had felt at the time. It had felt quite real. Everything looked clear and vivid, and I was able to think and speak quite clearly."
In a New Scientist article in 2000, she again wrote of this:
"It was just over thirty years ago that I had the dramatic out-of-body experience that convinced me of the reality of psychic phenomena and launched me on a crusade to show those closed-minded scientists that consciousness could reach beyond the body and that death was not the end. Just a few years of careful experiments changed all that. I found no psychic phenomena—only wishful thinking, self-deception, experimental error and, occasionally, fraud. I became a sceptic.[10][11]"
* She is a Fellow of the Committee for Skeptical Inquiry (formerly CSICOP)[12] and in 1991, was awarded the CSICOP Distinguished Skeptic Award.[4]
* Susan Blackmore has made contributions to the field of memetics.[19] The term meme was coined by Richard Dawkins in his 1976 book The Selfish Gene. In his foreword to Blackmore's book The Meme Machine (1999), Dawkins said, "Any theory deserves to be given its best shot, and that is what Susan Blackmore has given the theory of the meme."[20] Other treatments of memes, that cite Blackmore, can be found in the works of Robert Aunger: The Electric Meme,[21] and Jonathan Whitty: A Memetic Paradigm of Project Management.[22]
This is all from her wikipedia, which is about a one minute read.
As I recall, her consciousness textbook is fairly opinionated but it still gives a fair tour of the mainstream theories. She happens to be an illusionist like Dan Dennett. There's no "woo" in her coverage of the topic.
People who like to talk about consciousness have a problem keeping the goalposts in place (and trying to exclude cognition from consciousness, presumably because "cognition" doesn't sound mysterious enough, is a great example of this sort of moving of goalposts).
The truth is consciousness is an emergent phenomenon, and it can only be understood as such. Of course we have a lot more to learn, but we understand the fundamentals of neuroscience and biochemistry, and we know how our conscious experiences emerge from those. A neuroanatomy textbook (eg.: Blumenfeld) is probably the best place to start. There is unfortunately no single "thing" called "consciousness" that we will ever be able to point to.
> Consciousness" is not a thing. It's a collection of things, […]
A collection of things is a thing.
> The truth is consciousness is an emergent phenomenon, […]
I don’t think that is controversial at all.
We have unraveled the workings of many emergent phenomena. Emergent doesn’t mean incomprehensible.
In programmerland, one might say something like "the abstraction you're using is counter-productive", for whatever reason, when looking at some junior hire's code. It's very much the same here: trying to understand consciousness is asking the wrong question. It's like asking "Why is there a God?": it's an unproductive question because it's a question that can't be answered. The abstraction in the question simply doesn't map 1:1 to abstractions in the world in which we live. If we want to understand our world, we'll maximize our odds of success if we focus on studying the abstractions of this world.
I wouldn't describe myself as a pessimist at all, and I'm not sure why you thought my comment was pessimistic... But this opinion of mine (because that's what it is, it's an opinion, not an argument) is a professional opinion. This isn't my field of expertise, but I do have expertise in very adjacent fields.
And a pattern is a thing.
Consciousness is an everyday experience that appears to be reliably shared by billions of entities who communicate about it to each other. And have done so for at least millennia.
To use a colloquialism: It’s a thing
Replace type with concept and value with thing, and we get a rebuttal using the same language. You can also switch the terms, depending on how 'a thing' is interpreted. If you are more familiar with set theory, replace type with proposition, value with element. The critical difference is between reference and referent.
Further if large numbers of people holding something is true for considerable duration is sufficient for something to 'be a thing', we'd also have an argument for 'the consciousness of the human heart' being a thing, which I don't think anyone would argue in good faith these days. There are probably better examples, but note that the duration of time and frequency of experience are parameters with no obvious values indicating at which point the underlying 'thing' should be accepted as undeniably 'a thing'.
The whole point of requiring it to be technically defined is that all this confusion just goes away and it becomes possible to ask more meaningful and less ambiguous questions about it.
That said, I will concede that the argument I have made does not say that consciousness is definitely not 'a thing', just that it is defensible to say that it is not a thing, given that it is held that either it does not represent a proposition, or that the proposition it represents is not true. (note I'm using the term proposition with some handwaving).
The most fundamental aspects of reality, and mathematics, are still not founded completely on formal definitions, without resorting to circular reasoning, and attempts to prove things assuming the things that are being proved.
Existence isn't well defined. Yet Descartes' famous informal claim "I think, therefore I am" is still a compelling statement that each of us can use to verify we do exist. Despite many unanswered questions.
Likewise, "I am aware of myself, my own thoughts, my interactions with a world around me" is a compelling statement supporting consciousness. The fact that we can all verify our own consciousness the same way, and discuss those experiences with each other, makes it even more compelling.
Despite lots of unanswered questions.
--
Your counter examples are also weak. Most subjective beliefs about hearts having consciousness are not nearly as credible as informal arguments for existence and consciousness.
Conscious hearts are not virtually universally experienced.
Informal reasoning about why a heart, a blood pumping organ, would correlate with consciousness are not generally accepted informal arguments.
People whose hearts are replaced with artificial or donated organs don't reliably report a loss of consciousness.
My claim is that while tools to determine if consciousness exists do not presently exist, it is possible that they are discovered, and applied, and that the question of the existence of consciousness is open. Note that for me existence is a fundamentally formal notion. I cannot defend this except on grounds of its anecdotal usefulness because it is akin to an axiom or ground truth.
The notion of consciousness I consider is with respect to its 'actual' existence[1]. Suppose I imagine a circle, and then a square, but believe I imagined at once a round square. That I can consider the notion of 'a round square' or may have had a direct perception of it does not require me to accept its existence. Likewise, I do not need to accept the existence of a unifying singular notion of what might be perceived as consciousnesses. This also does not necessitate I reject it. If consciousness cannot find an agreed upon definition, assertions regarding it lack semantic meaning and cannot be 'reasonably' (waves hands) assigned truth values. If an exact definition is determined with no logical contradictions, I defer to the physical sciences for determining its plausibility.
The notion of conscious hearts was more credible in the past. The idea was that before the role of the heart was understood, a large population could have been convinced that 'the heart is the seat of consciousness', and thus a large population believing they were experiencing something a particular way may have been followed by the falsification of the existence of such a thing. At the time, it was very plausible that it was considered impossible to ascertain whether or not there were thinking hearts. I concede that the notion is not presently credible; for my purposes it is sufficient that it was ever credible.
If you regard existence proofs as being extra-logical, it may not be the case we can find agreement since we will be effectively talking in different languages. Personally, I am very skeptical of the ability of the human mind to find truth informally, which is why I lean on formal notions heavily.
--
1: My favorite essay on this distinction is Quine's (admittedly polemic) "On What There Is". The distinction I consider is slightly different from his in that I am also expressing distrust of direct self-perceptions of the mind's functions.
Just because someone believes something, doesn't mean there is a compelling argument it exists.
But if something can be studied by many independent people, with strong agreement not based on passed-on beliefs, then there is likely to be something there.
I.e. I am unaware of a universal experience of souls, or informal arguments for souls that are compelling.
I am aware of informal arguments for existence ("I am, therefore I exist") and consciousness ("I experience myself, my own thinking, even this thought itself, so I am conscious of myself"). In both cases, the vast (vast, vast) majority of humans conclude they do exist and are conscious. And the informal reasoning is very strong, even informally tautological.
Soul's don't have that kind of support.
However, I find bringing the concept of "soul" into the equation to be helpful, because many people intuitively believe in such things. If you ask them if souls are important, they'll say yes. If you ask them if something special happens to souls after death, they might say yes. But ask them to define what they mean when they say "soul", and you'll get a ton of hand-waving. People can believe very strongly in things they can't even describe or point to: this always puzzles me, because it shows that people don't even know what it is they believe in in the first place.
Whether the concept of "consciousness" is as ridiculous as the concept of a "soul" or not, I have found that they both suffer from the same definition problem.
There are many, many concepts which throughout History entire societies have shared a belief in, but which didn't map to real things that could be defined precisely. A "soul" is one of those, but other examples abound in the History of science and in the History of medicine.
Colors, sounds, tastes, smells, feels, emotions that makeup perception, imagination, inner dialog or visualization, dreams, hallucinations, illusions.
Or qualia if you want to reference the philosophy of mind literature, where the debate and various positions are well defined and extensively argued.
All of which are things we can explain without any magic.
The literature on qualia pretty much has the same problem as the literature on consciousness.
The "Why is red RED?" thing is a complete joke from a scientific standpoint.
> Every aspect of your experience is your nervous system and the sensory organs at work
There's still a leap from that to the existence of subjective experience.
It's almost like by insisting that there is a subjective experience (as opposed to what?) you close your eyes to how that experience arises, and prefer instead to implicitly attribute it some magical attributes such that it can never be explained mechanically.
BUT, we are SO similar genetically that there is SHARED properties to nearly all experiences among ALL humans, strongly arguing for objectivity to it.
Why is red specifically what we perceive as red is not answerable. How it affects you cognitively and emotionally is based on the entirety of your upbringing and past experiences as well as your sense neurology. Our brain has quite a lot of neurons and connections! Trillions of connections and billions of neurons! A lot can change and be affected in innumerably complex ways! It's more than sufficient to explain the entirety of our subjective experience.
This control and access, which needs a true observer, is consciousness. If a computer is taken as layer of reality, then this control is in outside layer of reality. But we don't stop there: It is observer in all layers of reality, whatever reality is made of. The observer can never be represented in any layer of reality.
We should consider that experience layer, like sound and vision are an independent layer of reality than underlying "material" reality. A materialist will say latter exists a priory, and thus experience is off the material layer. But we can as well argue that experience layer is a-priori. My taste sense, for e.g. has no equivalence in material reality. It deserves its own layer of reality. If I hear a voice of someone, it is singular experience rather than experience of collection of frequencies. Color like white is an experience of multiple frequencies of light. Our perceptions, happening with help of neurons, are not based on light at all. Projection is a-priory present, from which we create even light and other formulations...
A mathematical reality is in its own layer of reality. A conceptual reality of say, chair, is also in a "mental" reality.
So we should assume that all are simply abstract layers of reality. But our mental conditioning has created "things" model, just because we can touch and feel, or perceive a boundary in what we see.
As observer, which creates memory of observation, and accesses it, all in the present. An observation which is not in time and space, but creates sense of time and space out of observation.
And so on... A totally different model of reality.
https://pubmed.ncbi.nlm.nih.gov/12613670/
There's a recent book called "Conscious: A Brief Guide to the Fundamental Mystery of the Mind" which purports to be what you are looking for. She's an author who is married to Sam Harris, a neuroscience PhD and public thinker who focused a lot of consciousness. His ideas permeate her thinking and I found her treatment to be less open minded.
There's a series mirrored on YouTube called "Closer to Truth", where Robert Lawrence Kuhn interviews famous physicists, mathematicians, etc. on this and other grand unsolved questions. You can find interviews with most of the analytical thinkers in the consciousness space if you wade through.
The most "mathematically rigorous" theory is Integrated Information Theory by Giulio Tononi. The logical conclusions of his theory are outlandish or eye-opening, depending on your bias. I don't buy it myself but the there might be something in the approach.
The most mathematically gifted individual with a theory is Roger Penrose. His scientific work is absolutely brilliant but his ideas about consciousness are quite frankly ridiculous and unmotivated.
The best, most rigorous theory I know of comes from a Neuroscientist at Princeton named Michael Graziano. He put forth the "Attention Schema Theory". I strongly recommend his book "Rethinking Consciousness" which makes a case for it, and provides an overview of the other approaches and how they relate to his ideas.
I've been very much swayed by Graziano's work, and we have the technology today to test his theory by integrating actively controlled attention mechanisms into machine learning models. If his ideas are correct, that could lead to models organically making similar claims to humans regarding phenomenological states. The theory is agnostic about whether the claims would be substantiated.
Altered States of Consciousness: Experiences Out of Time and Self (The MIT Press) https://a.co/d/ghPhlDW
This paper is a good, accessible overview of how this would work in neuroscience:
https://www.pnas.org/doi/10.1073/pnas.2116933119
Virtualism as a Perspective on Consciousness, https://www.youtube.com/watch?v=b6oekXIQ-LM
It honors the richness and complexity of conscious experience and provides a refreshing computational perspective on it.
Consciousness is thus the universe peering back into itself, and living systems are the biotechnology which animate this phenomena.
The unique properties of this are:
Our “computation” is non-linear. The quantum coherence creates a holographic subjective singularity. These allow the illusion of one perspective among billions of participating neurons.
The more neurons in a cluster, the higher resolution the “pocket of consciousness”. These bubbles compete in the mind for influence. I say this acts much like a hologram.
The sieve of consciousness works through constructive and destructive interference.
Your downvotes are your HYPOCRISY!
I have a whole long list of assertions you will fight me on for the love of your ignorance.
In this moment I cannot fully invest in exhaustive debate.
I will prattle on a little more in an unstructured form for the sake of intrigue.
There are several things going, one being this parallel discussion on quantum entanglement (and the parallel discussion for our minds being hackable for another time.)
I would like to describe to you the technology of consciousness in ways I have come to understand, yet I will expect to be scorned and badgered for the inherently abstract nature of our topic.
Neurons in isolation exhibit a telemetric awareness of one and other
The Cosmic Serpent was inspiring in my youth. The primitive ideas the author discusses relate.
Microtubul structures likely provide a substrate for our experiential being. Electrochemical processes flash these substrates, and then gently and delicately suss the state resolve as though a sieve + stack.
The sieve is the constructive /destructive aspect.
I say “stack” as resolve is apparently pushed and pulled or read/write like memories, yet I believe this is where our non-classical thought strategy arises. The outcome is more like a micro-inference rather than a faithful memory.
Constructive and destructive patterns form an analog isolated holographic “image” (though not merely a picture, in every shade and depth of every experience we are capable of having.)
Any information can be pushed into this stack, yet interference collapses to associations (or novelty) producing a primitive psycho-emotional echo, other parts of the brain process into higher level responses.
Consciousness is not intelligence. Consciousness coopts intelligence as a tool for perpetuation.
How does this not sound like consciousness: “consciousness is the inflection upon the potential of our existential being.”
Existential being is any matter, yet all cold matter is dormant without animation through life or other technology.
The potential is the hyper-dimensional capacity of the carefully arranged atoms in our neural substrate.
Inflection is any perturbation of the quantum dynamic (which is usually pure chaos.) it is the atom in the universe which exists, the living system is a technology which extends and allows the aggregate scope of potential awareness and feedback.
The quantum layer is like an echo chamber for your jelly bag brain. This echo chamber is the material universe (which you are) echoing experience upon its analog subjectively scoped potentials.
See the idea, not the comment quality.
I always love to imagine how fun an exercise it must be for a writer to invent those dialogues.
A) We can't even define consciousness, let alone describe it's physical nature.
B) Even if it depends on quantum mechanics, it doesn't make it magic or unique to biology.
C) It's an extraordinary claim, so back it up, man.
I don't care about curious extrapolation, except in science fiction.
I DEFINE CONSCIOUSNESS AS THE INFLECTION UPON THE SUBSTRATE OF EXISTENTIAL BEING!
B) I say all life has some smidgen of this magic. It is non-classical and will not emerge out of complexity. All humans do not have “an equal amount”. Next argument may be of capacity.
The universe is the medium, consciousness the inflection, life the technology which animates.
C) voices in our heads. I know, I know, you don’t want to believe that either. Pay attention, check back in ten years. Hear them? Doubt we’re being screwed with? That old “master mind theory” from a hundred years ago touched upon this before our taboo of speaking of these voices dismissed them as crazy. Snail telegraph? There are clues throughout history.
We’re entangled, and it’s more of a hassle than a gift.
The mathematical structure of phenomenology is kind of a weird specific task, because it's traditionally equated to the structure of the world or representation in general. This is because in theory 'consciousness' means very little when abstracted from the theory of its implementation, on the other hand if we try to preserve the relation of consciousness with the implementation we already have to choose between several. I think the best candidate for the existence of some of most rudimentary sentience is the gamma oscillations that function as coordinator functions of different neural populations, so the math to represent this would be take that as a basis.
I'm not going to waste my time expecting to find answers to such things on a message board for software engineers and money men. May as well be 4chan
Just putting this out there in case anyone has read this and would care to share an informed opinion with a layperson!
> term ‘aspect’ denotes an instantiated phenomenal property or quale. The set of aspects A(e),
> then, comprises the phenomenal properties or qualia which are instantiated in the experience e,
> also called the phenomenal states of the experience e. Our question, then, is what it means
> that “any set of phenomenal states of a subject at a time is phenomenally unified”
> [...]
> A promising answer is the
> so-called subsumptive unity thesis, developed in [5]:
> “For any set of phenomenal states of a subject at a time, the subject has a
> phenomenal state that subsumes each of the states in that set.” [5, p. 20]
> According to this thesis, what underlies the experience of the unity of a conscious experience
> is that for any set X of phenomenal states in the conscious experience, there is a further
> phenomenal state that subsumes each of the states in X. This phenomenal state characterizes
> what it is like to be in all of the states of X at once [5, p. 20].
> [...]
> Phenomenal unity gives rise to a mathematical structure of conscious experience.
When I try to state this in natural language I come to something like "Conscious states consist of qualities, or contain qualia. And they are unified _in_ consciousness, meaning that our experience is whole. The mathematical structure is this: For any set of experiencing qualities there is a(n additional) state integrating them all into one state."
When I studied philosophy in the 2000s we discussed such kinds of solutions already in a seminar and found them lacking on two grounds.
First: there is no clear notion of non-conscious qualia, and Ockham help us here to keep it so. But if this is true, each of the integrated states already presupposes the whole (because it is already conscious, consciousness is more the medium of qualia), which then cannot come after them to unify them. Unless one wants to assume that a lot of conscious splintered experiences are integrated into a whole later. That would be even more questionable, because who experiences them?
Second: The integrating state needs another integrating state to integrate it into the whole of conscious experience etc. ad inf. Unless it is a state that integrates itself into the whole as well, which would just be a sleigh of hand - because exactly that is the riddle that a mathematical structure must overcome to be about consciousness. The self-reference, the self-presentation of consciousness. Apologies for lacking the proper terminology here.
edit: formatting
I agree that "non-conscious qualia" is an incoherent notion, but I'm not convinced this rules out a "phenomenal unity" as you argue - they are positing the unity as "phenomenal" not as the substrate (or medium, as you put it) of qualia, so more than just the raw fact of consciousness is implied.
One point of comparison might be Wittgenstein's famous example of the duck-rabbit - a bistable ambiguous image that can be seen either as a picture of a duck or as a picture of a rabbit. One might assume the qualia associated with looking at the picture are the same whether one is seeing it as a duck or as a rabbit - the same shapes and colours - and yet there is a difference:
> The change of aspect. "But surely you would say the picture is altogether different now!" [...]
> If I saw the duck-rabbit as a rabbit, then I saw: these shapes and colours (I give them in detail) -- and I saw besides something like this: and here I point to a number of different pictures of rabbits [...]
> 'Seeing as ...' is not part of perception. And for that reason it is like seeing and again not like.
Perhaps this un-perceptual yet pervasive and arguably phenomenal unity subsuming the duck-rabbit qualia - that which flips from duck to rabbit and back - is analogous to what the authors intend with their "subsumptive unity thesis".
Congrats on commenting on TFA, btw :) I thought it was an interesting read, though I start to lose my grip on the math around the topology section.
Distinction: to distinguish anything from anything else. Consciousness: the distinction of self from other. Also has further distinctions of inside the self and the other organized into patterns, and distinctions drawn for negative entropy and spiritual evolution, the basis of all change. Distinctions have three parts, that which is distinguished, that from which it is distinguished, and the consciousness drawing the distinction. This has reality differntiation along the lines of the Perceptual, Conceptual and Existential, and distinctions themselves have variables of Intent, Content and Extent.
Imagine a idealist, unified monism 9-D model of reality that has three main prongs. Other unstable particles teased into existence temporarily by LHC are not very useful whereas normalization based on the characteristics of the free electron when it comes to Space/Time/Consciousness are relevant.
For symbolic representation of Extent: this is Space that has three dimensions and is represented using integers. And then there is Time, which are image events of cause and effect on a timeline, that are represented like a plane, a volume, or that of more other people, and this represented with imaginary numbers. And then there is Consciousness-awareness, complex numbers, the awareness of being, or the distinction of Self from Other as spiritual awareness or life trace, and it too is a plane, a volume, and that of more other people.
For the symmetric resonance of Content we have Mass (subatomic, atomic and the universe), Energy (photonic, electric, magnetic) and also Consciousness-meaningfulness (individual, group and cosmic) aka gimmel, the third form of reality which is not mass or energy.
Primary Consciousness is one thing, but then gimmel, the organizer, is second and mass/energy is third.
For Intent or meaningful information: existential, conceptual and imaginary.
In this model we are all moving towards experiences like what is termed Cosmic Consciousness (omnipresence) in Yoga, especially along the lines of Kriya Yoga as brought forward into the West by Yogananda. Anyway, perhaps a start towards mathematizing consciousness that is compatible with ancient wisdom and modern mathematics.