Hacker News Reader

Comment by TheHideout | Hacker News Reader

Tangent by PhD student in Cognitive Science here - My current belief is that consciousness arises at the cross section of 3 core skills: Theory of Mind [0], Autobiographic Episodic Memory [1], and Embodied Cognition [2].

Principally, recognizing that other people have intentions and beliefs different from your own, being able to understand that you have experienced a timeline of personal experiences and these contribute to who you are, and that you are a physical being that can think in-situ as well as plan your physical actions in the future.

I'm favoriting this link and will try to return to it after my dissertation is done for reflection.

[0] https://en.wikipedia.org/wiki/Theory_of_mind [1] https://en.wikipedia.org/wiki/Episodic_memory [2] https://en.wikipedia.org/wiki/Embodied_cognition

View on HN

There are actually two problems, that people often mix up. One is how we get a sense of self, which I think is what you're talking about. Another is how qualia, like the perception of color, arise at all. Nobody seems to have any idea how that happens; we have no way to map some arrangement of atoms to a perception and prove from first principles that one leads to another.

And the second problem (the hard one) is the one the linked article is focused on.

Franky, I have never understood the problem of qualia - or why it is something in need of explanation at all. Maybe the correct answer is an "anthropic principle"-style non-answer: External stimuli have to affect the human brain and body somehow. There has got to be some internal representation of what you percieve. Now we have the capability of reflection. To some extent, we can think about how we think, and we can perceive how we percieve. Naturally you then can wonder: Why/how does this stimuli lead to this internal state? But if the states were different, you would be wondering about those states. And you only wonder about it, because you can due to reflection.

"There has got to be some internal representation of what you percieve."

I don't think this is necessarily true. It seems to me like the universe would get on exactly the same without internal representation. By "internal representation" I mean a universe without "embodiment" -- no experience of being me or being you, no feelings felt, no colors seen, no sounds heard, etc.. And yet those experiences "exist" in some sense. Call them illusions or whatever, but I am and you are. So the problem is, if we can imagine the universe getting on exactly the same without those experiences, why is it such that I am and you are. That seems to be the puzzling question to me -- that nothing has to be experienced, and yet things are being experienced.

> So the problem is, if we can imagine the universe getting on exactly the same without those experiences, why is it such that I am and you are.

The most plausible answer, in my opinion, is that our lack of knowledge about how our minds work means that we are free to imagine something that probably isn't so. Imagination is not a good enough guide to how things are to be the considered a basis of knowledge.

That's a bit of a jump from no internal representation to no embodiment (whatever embodiment means in this context as that is pretty unclear to me)

I take internal representation to be an aspect of "embodiment", which you could also call the experience of being in the first person. If some idea is being represented internally, then it is being experienced, and the sum total of all these different experiences happening together (seeing, hearing, thinking, feeling etc..) are what you could call "embodiment".

I think philosophy is a difficult subject partially because people use different terms, or sometimes their own unique terms, to describe subtle and complex ideas.

When I mentioned embodiment, I was referring to the experience of being in the first person while feeling hot or cold, seeing colors and objects, hearing the sound your keyboard makes when you press keys, feeling the keys with the tips of your fingers, etc. All of that together at once.

1) The issue of what perceptions "actually" are will be impossible to answer because qualia are ireductible.

Let me explain: you feel pain in your foot.

As a scientist I cut your foot and you no longer feel pain in your foot.

Or I cut the pain nerves, or the spinal cord, or give you pain medicine.

Or I find a brain receptor and a chemical responsible for pain.

Or in the future, we find a couple of brain networks that when hacked turn of or exacerbate this and that sensation. We find the exact neuron microstructure that is actually a quantum computer and a certain process/instruction set can precisely manipulate sensations. We find the exact process that is responsible for the "self" sensation.

I know that would still not satisfy my question of "what are sensations?".

"Yeah, this and that neural network, but WHO is actually feeling that sensation?".

It would be a little bit like trying to "see" in 4D space -- impossible actually.

2) Why are sensations really needed?

If we get to that advanced level as to technically answer the first question, even without a gut feeling understanding, I think we could actually understand the answer to this question.

The answer might be -- "because AI can happily process numbers, but the biological quantum neural nets cannot, and they need this "sensation" representation to actually work in the most efficient way possible.

By "quantum" I mean some missing stuff we don't know yet, I'm not saying the brain is a "quantum computer".

> But WHO is actually feeling that sensation?

This reminds me of a story from Jon Bentley's Programming Pearls. An early multiuser computer system did fine with a small number of users, but choked when one more was added, above a certain threshold. "No problem", said some wag "just find where that threshold is stored and increase it." The joke, of course, is that the number didn't exist as an explicitly-realized value; it emerged as a phenomenon out of how the system as a whole ran.

When you are in pain, your conscious mind - an emergent phenomenon from the physical processes of your body (primarily the brain), or so I suspect - observes itself. There is nothing mysterious in systems observing themselves; that's what a computer doing when it raises and responds to a segfault, for example.

I'm not saying it is incorrect, but the emergent theory of self-awareness is a bit hand-wavy.

Surely a primitive man could say an airplane flying is an emergent property of the airpland, like Aristotle said about many things?

I was just questioning the level of satisfaction we would get as science would reach the best explanation of what sensations are.

I think right now we are nowhere near that scientific knowledge.

Considerable vagueness is to be expected in any field so far from an explanation, and the situation is even worse in the dualist camp: to the best of my knowledge, they have never been able to offer one definite, affirmative claim about how the alleged non-physical aspects of the mind work. All they do is weave tendentious arguments against there ever being a physical explanation of the mind.

Emergent phenomena are pervasive in complex systems, and this stands as an effective response to naive dualist misconceptions such as that materialism requires that a quale must be identical with a single physical thing.

Even in airplanes, behavioral traits such as stability emerge from the interaction of several physical features.

> I don't think this is necessarily true.

Most people, say 99% or so, can form an internally perceived image of something in their mind's eye. The rest aren't seeing things in their mind's eye, other than that which is built from direct sensory input. Both are form.

  You see with your eyes
  I see destruction and demise (that's right)
  Corruption in disguise
  From this fuckin' enterprise
  Now I'm sucked into your lies
  Through Russel, not his muscles but percussion he provides
  For me as a guide
  Y'all can see me now 'cause you don't see with your eye
  You perceive with your mind
  That's the inner
  So I'ma stick around with Russ' and be a mentor
  Bust a few rhymes so motherfuckers remember where the thought is
  I brought all this
  So you can survive when law is lawless (right here)
  Feelings, sensations that you thought was dead
  No squealing, remember that it's all in your head

The problem with qualia, as I understand it, is that we can build machines that process images, and in the future let's say, can exceed what cat is capable of, but that machine will never actually have sensations.

So there are two issues: what sensations actually are and why was it necessary for life to have those sensations, after all, a robot cat would do just fine without any sensation, just processing the inputs and reacting to those inputs.

You say the machine wouldn't have sensations. Yet a cat and a human also are just complex machines. Maybe being a sufficiently complex machine is all that is needed to have sensations.

And if we could imagine for example that an ant or a fly actually experiences things, if it too has some sort of degree of what we as humans have when we see or hear things, maybe there actually already are machines that "see" and "experience" things.

The difference might just be that we as humans have the extra layer of meta observation that allows us to reason about our experience in the context of the world and turn us into an active participant in that experience while an ant or a neural network are mostly just passive observers even though they too make decisions and process information.

Complex is a very tricky question.

I think people already simulated a worm's neural nets, but we simulated just high level abstractions of what those "simple", 200 neuron networks are. Sure, by simulation tickling part of the worm, we get the exact reflex reactions a living worm would, but surely the simulated worm does not have sensations, I can do the same worm simulation with if-else statements, just because I see a worm realistically wiggling on the screen I do not believe it can suffer pain, or else I would never play Worms Armageddon.

I believe there is something to the chemical processes themselves that we will not be able to replicate in pure software.

That brings the question, what if we could simulate the quantum particles themselves, not just some high level representation of cells, would that simulated being have sensations/feelings?

If we perfectly simulated, at the quantum particle level a cat, would the cat behave like a real cat?

That is mindboggling to me; unfortunately we cannot do that yet.

I've sometimes wondered if people precieve things qualitatively different way then I do to give rise to this problem, since to me it seems like s non-problem. To me, i have trouble even conceptualizing what is meant by the qualia problem. What is the fundamental difference between something's "redness" and say a tall person's "tallness". Maybe "redness" is a bit more mysterious because its non-spatial, but still is it really that different of a quality it needs an explanation? What would even be explained in the explanation?

Tallness is a physical property of the thing itself.

Light also has its own physical properties, e.g. a wavelength of 700nm. But the qualia of red is the experience you have when that 700nm light hits your retina.

It's conceivable that when 700nm light hits my retina, I have an experience which you would call "blue." That's why the qualia and the physical property of the light are two different things.

We have a lot of evidence of how people perceive color, and the differences of perception. There are tests that measure the ability to distinguish color. Some people are only colorblind in one eye, and report perceptual differences that are predictable by phenotype. I would expect any meaningful difference in qualia would have some conceivable (if not performable) test. Otherwise we're calling two equivalent things different for no reason at all.

I personally tend to believe that non-falsifiable statements are necessarily non-physical.

I challenge you to devise a test able to detect that I perceive the color spectrum in reverse.

Is that a meaningful difference in qualia? Does a relabeling of what are arbitrary labelings change my subjective experience. Is precieving the colour spectrum in reverse even a meaningful statement? Or is it more like saying you can't prove i added 2+1 instead of 1+2.

Well, that's my point. If you can't detect it, it's not a difference.

But you dont experience tallness as an objective property of the object. You view it relative to the scene, people who intimidate you appear taller, etc. The tallness might cause us different emotional experience too - maybe im intimidated, maybe it makes me feel safe, maybe it makes me feel trusting, etc

Its entirely possible that when the exact same light hits our eyes to form a scene, you and i would experience the tallness of the person in the scene differently.

“ Another way of defining qualia is as "raw feels". A raw feel is a perception in and of itself, considered entirely in isolation from any effect it might have on behavior and behavioral disposition. In contrast, a cooked feel is that perception seen as existing in terms of its effects. For example, the perception of the taste of wine is an ineffable, raw feel, while the experience of warmth or bitterness caused by that taste of wine would be a cooked feel. Cooked feels are not qualia.” — Wikipedia

Qualia are easier to understand when you think about things you can't experience. For example, think about echolocation. We understand how it works, we understand what parts of a bat's brain light up when echolocation is used. However, I have no idea what the experience of echolocation might be like.

I guess it's kind of like hearing.. but different? Maybe a weird combination of hearing + seeing black and white? Or similar to one of those LIDAR depth-field renderings? In other words, I don't have access to the qualia of some echolocated object. Exactly how I wouldn't have access to the qualia of some red object if I were blind.

Why does an internal state lead to an experience of color or pain? Computers have internal state without any expectation of them being conscious, and this internal state can be the result of external stimulus like a camera. But the internal state isn’t experienced as color.

> But the internal state isn’t experienced as color.

How would a conscious entity experience an internal state in a way that's not like color (or pain, etc)? I think the claim is that conscious entities must experience some internal states somehow (otherwise what are you concious of) - how else could they do that other than via sensation?

I don't think that internal states of non conscious entities has much bearing on what a conscious entity does or is.

There has to be an internal representation for the organism to function properly. E.g. if light with a 450nm wavelength hits your retina, certain neurons fire. You get a pattern of electrical impulses. There's your internal representation.

But why does that representation also carry with it an experience of the color blue? Why isn't it just the electrical impulses and that's it? The organism would still work. We could be like robots, behaving exactly the same as our nerve impulses move us around, with no inner experience at all.

And what about electrical impulses in things other than brains? Are there qualia associated with them? Is there something it's like to be a computer, light bulb, thermostat, lightning bolt, or rock? If not, why not?

If you and I both see that 450nm light, how do you know that I'm not having an experience of color that you have in response to 700nm, i.e. red? Maybe the rainbow of colors is upside-down in my experience compared to yours. Maybe my qualia are different entirely. We have no way to prove anything different, because all we can measure is the electrical impulses. And because we haven't the faintest hint of a theory that connects the physical stuff to the particulars of the conscious experience.

> You get a pattern of electrical impulses.

Is there something else in the brain that's not just protein cascades and electrical impulses? I'd be interested in a published paper describing this structure.

Barring that, yes, experience really is just protein cascades and electrical impulses. A light bulb is orders of magnitude simpler — easily 10 or 12, maybe more; and, that's just spatial; the spatiotemporal complexity might be another 10 or 12 orders of magnitude. I don't think it's any stretch of the imagination to believe qualia arise somewhere between "uniform black body emitter" and "human brain learning and experiencing over time".

That's often people's assumption, that at some level of complexity, qualia just show up. But we don't actually have a clue why that might be the case, or how it happens, or why a pile of atoms banging against their neighbors should be considered a larger whole which thus has complexity and therefore consciousness.

People think qualia must just show up this way because if you're committed to materialism, it seems there's no other option. But that's not an argument, it's just faith in materialism.

And it's not true that there are no options besides simple materialism and crazy woo. E.g. it's possible that all matter has some level of conscious experience; you can believe that without believing anything contrary to physics. There are all sorts of possibilities being explored by people who grapple with these questions, instead of just assuming an answer.

> E.g. it's possible that all matter has some level of conscious experience

I've tried to reason about this, and one hypothesis is that all matter has compulsions, like to get to lower gravitational potential energy, to react this way or that, etc. And that desire can be thought of as conciousness. So for more complex assemblies of matter, conciousness becomes the sum of those compulsions, but still reflects the sum of what the base components want to do - oxidize stuff, etc. But the point is, conciousness is the manifestation of the inherent desires we have, based on the rules of the universe.

Just a hypothesis of course, not testable, but my demonstration is that if you pour water into a maze with an outlet, it will solve it, based on its desire to reach a lower potential energy, and to make hydrogen bonds.

> People think qualia must just show up this way because if you're committed to materialism, it seems there's no other option. But that's not an argument, it's just faith in materialism.

What other explanation is there? A magic fairy unicorn in the sky is making the qualia possible? How does that solve the problem you are pointing to, wouldn't we still want to get there via "first principals" ?

And if we did find a magic unicorn in the sky, we'd study it and probably eventually understand how it works, what forces are acting upon it or vice versa. And once we'd done all that, someone on a forum would say of course this can't explain qualia because its "just materialism".

If you start at first principles you don't find support for a fundamentally material reality. Idealist monism is the alternative.

i.e. the universe and everything in it is fundamentally mental phenomena (qualia) which may only appear to have material properties by the extent of their disenfranchisement from more connected phenomenal clusters.

Right. See Bernard Kastrup's work.

Or, Plato. Or Max Tegmark.

In my experience, most people can't follow their arguments because their woo detector turns on and their rational faculty turns off.

It's not much different than the magic unicorn in the sky that makes fermions pop into existence out of thin air (our current understanding of fundamental physics postulates just that). Every ontology gets a free miracle, you can't reduce everything in terms of something else, either you'll never stop or you'll get into circular reasoning. You get to choose your ontological primitive. A "particle" or "consciousness" do not have different woo-factor. But by choosing consciousness as a primitive you can explain much more of what we observe in nature (and as a nice side effect, there might be positive existential implications with such a worldview).

I like that idea that at a deeper level quantum computing matters in the brain. Add a few digits to the layers of complexity.

P I’m a Bnb

Something that might help laypeople like myself, would be to relate them to the colors of things. That's the third dimension in this puzzle. There are wavelengths, and we have nerve impulses, but in addition, things interact with wavelengths in different ways.

Talking about wavelengths and electrical impulses seems too abstract. Let's talk about a tomato. Do you experience a 450 nm (or maybe 550 to be fair) tomato the way I experience a 700 nm tomato?

Do we have qualia for empty and full? Do you experience an empty beer bottle the way I experience a full beer bottle?

If you consider a brain as a sparsely connected graph then it also explains these experiences. The nodes nearby to the color blue perception nodes are those that we think of as clearly related (sky, water, certain crayons), but also those that are less directly related: water, cold, sadness.

Is the question about qualia even a meaningful one? It feels to me like running a program twice and asking whether it's really the same program that was run, given that each time, the addresses of library functions the program used were different[0].

It could be that when you see 450nm light, the pathways in your brain that process this information correspond to the pathways that respond to 700nm light in my head. But it doesn't matter; our experiences are adjusted to match, because they're synchronized, thanks to shared environment and language. In both of us, despite invoking different neuron groups, seeing a blue sky will evoke images of water, or Windows error codes, and not of roses or rust. If we talk, both of us will use the word "blue" to describe what we're experiencing.

Whether the rainbow is upside-down or not in your head doesn't seem to be a meaningful question, if through all imaginable ways to observe and communicate it, it seems you perceive it the same way as the rest of us.

[0] - https://en.wikipedia.org/wiki/Address_space_layout_randomiza...

I would say the question of qualia is meaningful, in the sense that a good answer might teach us something about the nature of reality. It's just not a practical question.

However, it might become a practical question, if someday we have to decide whether AGI should have rights, or whether we think a human uploaded to silicon would have qualia.

Maybe qualia is generational knowledge; Certain distinct objects developed differentiation and over time wee learned to associate it visually with color. Initially in our evolution we only had two wavelengths of color differentiation, but somewhere along the line a mutation caused a third, offset wavelength to appear. This is what color blindness is - my third mutated wavelength is just slightly closer to the second wavelength.

Thanks for sharing. My own pet theory is very similar to yours: humans evolved consciousness in order to ‘observe’ our cognitive and emotional state, so that we can build models and apply them to reason about other people’s cognitive and emotional states. It’s hard to have ‘theory of mind’ about someone else when you don’t have access to your own mind to start with.

FYI, there's a paper that claims exactly this: https://pubmed.ncbi.nlm.nih.gov/20658859/

Thanks!

This sounds similar to claims I've heard that consciousness arises from language. Taking a component of our cognition which is heavily intertwined with the rest of it and taking its ubiquity to imply some sort of primacy.

Have you ever had a thought you can't put into words? Some people do not even possess an internal monologue!

That doesn't sound right to me at all.

To your first point: consider a hypothetical person who is unable to recognise other people as being similar to themselves, i.e. as having minds of their own. We would treat this person as having a serious social disability. We certainly wouldn't conclude that the person lacks consciousness entirely. They could still be capable of experiencing pain and pleasure, and could still be capable of seeking goals, and of abstract thought. They would still have moral standing and personal dignity - it would be immoral to mistreat them, and we would consider that human rights still apply to them.

We are deeply social animals, and this is integral to our sense of self as we usually see it, but I'm not convinced it's integral to consciousness.

I have a similar objection to your second point. Lacking a sense of personal narrative and timeline would constitute a serious disability, but would not render a person unconscious. If you had total retrograde amnesia, you would still be conscious.

I see the same problem with your third point. I'm confident that, under the right concoction of psychotropics, my mental function could be degraded to the point that I would be unable to reason effectively about my future, or to make decisions, while still keeping me conscious. Someone suffering something akin to shell-shock might be unresponsive, but still conscious.

Ha! Under the influence of certain psychotropics, your ability to do functional planning would be diminished, but the number of creatures that you thought worthy of your empathy might well grow. :-)

What would constitute "unconscious"?

Having your performance degraded to the level of a chaotic, finite state machine? (Or is one of those also conscious?)

Inability to pass the Turing test? (But then "unconscious" [???] machines can pass that.)

Consciousness is such a meme.

> To your first point: consider a hypothetical person who is unable to recognise other people as being similar to themselves, i.e. as having minds of their own. We would treat this person as having a serious social disability.

It sounds like you're describing psychopaths.

I suppose so, but I imagine psychopaths can still understand it at the intellectual level. Not something I know much about.

Thoughts about Epiphenominal explanations of the mind?

I'm a former philosophy student that had general interests Philo of Mind and that always stuck out to me as the most reasonable explanation for consciousness.

As a former psychology student (lol) I have to point out that an epiphenominal mind is perhaps the least parsimonious thing imaginable. It's also been effectively disproven since the cognitive revolution at least. Since we can understand ideas, take in / generate perceptions, and act in meaningful ways based on what we've understood - some kind of raw behavioural conditioning is not a valid explanation for our behavior. Nor the behavior of most higher animals.

>t's also been effectively disproven since the cognitive revolution at least.

How so?

>Since we can understand ideas, take in / generate perceptions, and act in meaningful ways based on what we've understood

That's not necessarily 'proof' of consciousness or free will.

>some kind of raw behavioural conditioning is not a valid explanation for our behavior. Nor the behavior of most higher animals.

Behaviorism isn't epiphenomenalism. I'm concerned you might not understand what epiphenomenalism is.

Can you elaborate on why you think epiphenominalism is the most reasonable explanation for consciousness?

My point of confusion with the epiphenominal view point is if mental states aren't at all causal, why should they exist at all? It seems like a clever way of compromising on the ontological question ("your mind is _kind of_ real"), but if mental states can't _do_ anything, then why should physical life develop them?

>then why should physical life develop them?

The comparison I've always liked is the "steam whistle set off by a locomotive". The steam whistle has no causal efficacy on the locomotive but is a waste byproduct of a steam engine.

Consciousness shouldn't have any causal effect on physical processes (assuming a strictly materialist framework), if it's explained as a byproduct of functioning physical processes it tends to deal with the issues 'consciousness' has that are in inherently in conflict with strict materialism.

I'm still confused. The steam whistle is built intentionally, alongside the locomotive, and together the train serves a clear, externally imposed purpose. The whistle doesn't impact the function on the engine, but a train with a whistle was decided to be a better train, because it does casually interact with the train's environment.

I get the ontological claim that mental states are real but cannot cause effects on physical states. But it seems unsatisfactory to claim that these rich, real but non causal states just happen to arise, without purpose, in only some physical systems, without further explanation.

I like this a lot, plus it’s a trinity :p

Please forgive me for Going There, but I believe [2] is also Original Sin, even in a secular definition as a generic source of evil perpetrated by humans against humans.

how will you test your belief?

While not a test for the belief directly nor a test of consciousness, I would start with making a sufficiently advanced embodied reinforcement learning agent that can do each of these things on a super basic level, then test for emergent goal setting and seeking.

That would be an interesting exercise in computer science, but wouldn't show us that you'd correctly determined the criteria for consciousness.

I absolutely agree. It might open some doors for further testing though.

Reply on news.ycombinator.com